| {:,:,:,:4,:dataset distillation\dataset condensation\,:,:,:,:State-Gated Retrieval\Only include arXiv records where v1 submitted falls between 2019-01-01 and 2021-12-31.\The title or abstract must explicitly use a phrase from the \\\ or \\\ family, and the semantics must refer to replacing large-scale training sets with a small number of synthetic training data points.\The current primary category must be cs.LG. Do not include records that are only cross-listed into Machine Learning by other primary-category papers.\The paper must have a version update in the calendar year following its initial submission, and the current record must still clearly indicate a subsequent publication venue.\dependency_type\Data + Control\intra_chain\inter_chain\data_dependency\Root searches yield candidate arXiv IDs\Each candidate abstract page yields primary category, v1 date, first next-year version, cue location, and publication-trail source\Answer rows are produced only after deduplication and detail-page validation\control_dependency\Records must be filtered by original submission year, with per-record validation that a next-calendar-year version exists\Cross-listed-only hits must be excluded, and the current primary category must be validated on abstract pages\Searches must cover both Title and Abstract because valid distillation hits can appear only in the abstract\Topic matches without a current publication trail still fail until the Comments, Journal reference, or Related DOI field is checked on the record page\freeze\historical_window\v1 year in {2019, 2020, 2021}\answer_type\multi-row ordered table\,:inclusion_conditions\Only include arXiv records where v1 submitted falls between 2019-01-01 and 2021-12-31.\The title or abstract must explicitly use a phrase from the \\\ or \\\ family, and the semantics must refer to replacing large-scale training sets with a small number of synthetic training data points.\The current primary category must be cs.LG. Do not include records that are only cross-listed into Machine Learning by other primary-category papers.\The paper must have a version update in the calendar year following its initial submission, and the current record must still clearly indicate a subsequent publication venue.\exclusion_conditions\Exclude records whose current primary category is not cs.LG and are only cross-listed to cs.LG.\Exclude records where the title or abstract contains the words \\\ or \\\ but does not correspond to the dataset distillation or dataset condensation research line.\Exclude records that have no next-calendar-year version update, or whose current record contains no clearly visible publication clue.\normalization\field_separator\|\record_separator\\\n\header_row\required literal first row shown in the oracle\schema\arXiv ID\term family\cue location\v1 submitted\first next-year version\date of that version\primary category\publication-trail source\current publication clue\dedup_key\bare arXiv ID without version suffix\date_format\YYYY-MM-DD\term_family_extraction\allowed_families\Distillation\Condensation\tie_break\if both families appear, prefer the title hit; if both appear only in the abstract, keep the first abstract occurrence\cue_location\Title\qualifying cue appears only in the title\Abstract-only\qualifying cue appears only in the abstract\Title + Abstract\qualifying cue appears in both title and abstract, or title carries only a generic cue while the abstract carries the specific family name\publication_trail_source\Comments\Journal reference\Related DOI\current_publication_clue\keep the shortest phrase that still preserves venue or publication-status information; when sourced from Related DOI, drop the DOI prefix and keep only the informative clue\sorting_or_selection\primary\v1 submitted ascending\secondary\arXiv ID ascending\stop_condition\all four corrected result pages are exhausted\all unique candidate IDs are opened to abstract-page depth\for every surviving candidate, the abstract page, current browse context, and publication-trail fields are checked before emitting the final ordered table\,:} |
| {:,:,:,:10,:,:,:,:,:State-Gated Retrieval\Only include arXiv papers whose v1 was submitted in 2022.\The title or abstract must explicitly contain clues related to chain-of-thought or self-consistency.\The current primary category must be cs.CL; do not include records that are only cross-listed from other primary categories.\Must have had a version update in 2023, and the current comments field must clearly indicate subsequent publication information.\dependency_type\Data + Control\intra_chain\inter_chain\data_dependency\root searches yield candidate arXiv IDs\each candidate abstract page yields primary category, v1 date, first 2023 version, comments/publication clue\answer rows are produced only after deduplication and detail-page validation\control_dependency\searches must cover both Title and Abstract fields because valid rows can appear only in the abstract\cross-listed hits whose primary category is not cs.CL must be excluded, and the primary category must be validated on the abstract page\topic-matched hits must be validated against a current Comments publication clue before they can be accepted\2023 candidates must be identified by original submission year and full version-history checks, rather than by latest-date proxies\freeze\historical_window\v1 submitted in 2022; version-history event in 2023\answer_type\multi-row ordered table\,:inclusion_conditions\Only include arXiv papers whose v1 was submitted in 2022.\The title or abstract must explicitly contain clues related to chain-of-thought or self-consistency.\The current primary category must be cs.CL; do not include records that are only cross-listed from other primary categories.\Must have had a version update in 2023, and the current comments field must clearly indicate subsequent publication information.\exclusion_conditions\Exclude records whose current primary category is not cs.CL, even if they appear in relevant search results.\Exclude records that contain only vague abbreviations or weakly related hints, but whose title and abstract lack explicit CoT/self-consistency semantics.\Exclude records that have no 2023 version update, or whose current record lacks a direct publication clue.\normalization\field_separator\|\record_separator\\\n\header_row\required literal first row shown in the oracle\schema\arXiv ID\cue family\cue location\v1 submitted\first 2023 version\date of that version\primary category\publication clue\dedup_key\bare arXiv ID without version suffix\date_format\YYYY-MM-DD\cue_family\Chain-of-thought\Self-consistency\cue_location\Title\qualifying cue appears only in the title\Abstract-only\qualifying cue appears only in the abstract\Title + Abstract\qualifying cue appears in both title and abstract, or title carries only a generic cue while the abstract carries the specific family name\publication_clue\source\Comments field only\extraction\keep the shortest phrase that still preserves the current publication-status clue\sorting_or_selection\primary\date of the earliest 2023 version ascending\secondary\arXiv ID ascending\stop_condition\all corrected result pages are exhausted\all unique candidate IDs are opened to abstract-page depth\for every surviving candidate, the current browse context, Comments field, and full Submission history are checked before emitting the final ordered table\,:} |
| {:,:,:,:5,:,:,:,:,:State-Gated Retrieval\Only include arXiv papers whose v1 submitted date falls in Q4 2020.\The title or abstract must explicitly contain vision- or image-domain Transformer-related expressions.\The current primary category must be cs.CV.\The paper must have had a version update in 2021, and the current record must clearly indicate a subsequent publication venue.\dependency_type\Data + Control\intra_chain\inter_chain\data_dependency\root searches yield candidate arXiv IDs\candidate abstract pages yield v1 submitted date, first 2021 version, primary category, cue location, and publication-trail source\answer rows depend on deduplication plus detail-page validation\control_dependency\searches must cover both Title and Abstract because valid positives can appear only in the abstract\candidate eligibility must be determined by the original-submission date window rather than by arXiv-ID or announcement intuition\candidate filtering must use full Submission-history validation rather than most-recent-submission filtering\searches must use the broader semantic cue state rather than a ViT-only entry point, because ViT-only hits can be false positives\publication-trail false positives still require Comments / Journal reference / Related DOI validation before final inclusion\freeze\historical_window\v1 submitted from 2020-10-01 through 2020-12-31; version-history event in 2021\answer_type\multi-row ordered table\,:inclusion_conditions\Only include arXiv papers whose v1 submitted date falls in Q4 2020.\The title or abstract must explicitly contain vision- or image-domain Transformer-related expressions.\The current primary category must be cs.CV.\The paper must have had a version update in 2021, and the current record must clearly indicate a subsequent publication venue.\exclusion_conditions\Exclude records whose current primary category is not cs.CV and are only cross-listed.\Exclude records that contain only weak cues such as Transformer or ViT, but whose title and abstract lack explicit vision- or image-domain expressions.\Exclude records that had no version update in 2021, or whose current record contains no direct publication clue.\normalization\field_separator\|\record_separator\\\n\header_row\required literal first row shown in the oracle\schema\arXiv ID\cue family\cue location\v1 submitted\first 2021 version\date of that version\primary category\publication-trail source\current publication clue\dedup_key\bare arXiv ID without version suffix\date_format\YYYY-MM-DD\cue_family_extraction\record every qualifying family actually present and join multiple families with `; ` in the emitted field\cue_location\Title\qualifying cue appears only in the title\Abstract-only\qualifying cue appears only in the abstract\Title + Abstract\qualifying cue appears in both title and abstract, or title carries only a generic cue while the abstract carries the specific family name\publication_trail_source\Comments\Journal reference\Related DOI\current_publication_clue\keep the shortest phrase that still preserves venue or publication-status information; when sourced from Related DOI, drop the DOI prefix and keep only the informative clue\sorting_or_selection\primary\date of the earliest 2021 version ascending\secondary\arXiv ID ascending\stop_condition\all corrected result pages are exhausted\all unique candidate IDs are opened to abstract-page depth\for every surviving candidate, the abstract page, current browse context, Comments field, Journal reference, Related DOI, and full Submission history are checked before emitting the final ordered table\,:} |
| {:,:,:,:5,:,:,:,:,:State-Gated Retrieval\Only include arXiv papers whose v1 submission date falls between 2020-01-01 and 2020-06-30.\The title or abstract must explicitly focus on randomized smoothing, de-randomized smoothing, transformation-specific smoothing, smoothed classifiers, certified defenses against backdoor or poisoning attacks, or partition aggregation.\The current primary category must be cs.LG.\The paper must have had a version update in 2021, and the current record must clearly indicate formal publication or acceptance information.\dependency_type\Data + Control\intra_chain\inter_chain\data_dependency\root searches yield candidate arXiv IDs\candidate abstract pages yield v1 date, primary category, full Submission history, cue location, and current publication clue\final rows depend on deduplication plus detail-page validation\control_dependency\candidate filtering must use the original-submission window together with full 2021 version-history validation, because valid rows can have later newest versions\cross-listed hits whose primary category is not the target category must be excluded, with category validation on detail pages\searches must cover the broader certified-defense family, including partition-aggregation and backdoor/poisoning variants, rather than smoothing-only entries\topic+venue hits must be confirmed by detail-history validation that a 2021 version actually exists\freeze\historical_window\v1 submitted from 2020-01-01 through 2020-06-30; version-history event in 2021\answer_type\multi-row ordered table\,:inclusion_conditions\Only include arXiv papers whose v1 submission date falls between 2020-01-01 and 2020-06-30.\The title or abstract must explicitly focus on randomized smoothing, de-randomized smoothing, transformation-specific smoothing, smoothed classifiers, certified defenses against backdoor or poisoning attacks, or partition aggregation.\The current primary category must be cs.LG.\The paper must have had a version update in 2021, and the current record must clearly indicate formal publication or acceptance information.\exclusion_conditions\Exclude records whose current primary category is not cs.LG, even if semantically related.\Exclude records that are only broadly related to robustness, defense, or backdoor/poisoning but do not meet the required certified-defense semantics.\Exclude records that have no version update in 2021 or no current publication clue.\normalization\field_separator\|\record_separator\\\n\header_row\required literal first row shown in the oracle\schema\arXiv ID\defense family\threat/topic cue\cue location\v1 submitted\first 2021 version\date of that version\primary category\publication clue\dedup_key\bare arXiv ID without version suffix\date_format\YYYY-MM-DD\defense_family_extraction\record every qualifying defense family actually present and join multiple families with `; ` in the emitted field\threat_topic_cue\keep the shortest qualifying threat or topic cue that justifies inclusion\cue_location\Title\qualifying cue appears only in the title\Abstract-only\qualifying cue appears only in the abstract\Title + Abstract\qualifying cue appears in both title and abstract, or title carries only a generic cue while the abstract carries the specific family name\publication_clue\keep the shortest phrase that still preserves venue or publication-status information\sorting_or_selection\primary\date of the earliest 2021 version ascending\secondary\arXiv ID ascending\stop_condition\all corrected result pages are exhausted\all unique candidate IDs are opened to abstract-page depth\for every surviving candidate, the abstract page, current browse context, Comments field, and full Submission history are checked before emitting the final ordered table\,:} |
| {:,:,:,:20,:,:,:,:,:State-Gated Retrieval\Use only North Carolina county data from the 2014-2018 American Community Survey 5-year estimates, and use the North Carolina statewide values from the same release as the comparison baseline.\Retain only counties with at least 4,000 households lacking internet access, and where the share of households without internet access, the share of households with a householder aged 65+ living alone, and the share of households without a vehicle are all higher than the statewide shares, and the median household income is lower than the statewide median.\Output the final results sorted alphabetically by county name, retaining the six required fields.\dependency_type\Data + Control\intra_chain\inter_chain\data_dependency\the 2014-2018 ACS 5-year product and North Carolina county geography yield the full county candidate set plus the statewide baseline row\the relevant ACS tables yield no-internet households, senior-living-alone share, no-vehicle share, and median household income for each county and the state baseline\final rows are produced only after the 5-year product/geography setup is fixed consistently and the full county filtering chain is executed from B28002 through the downstream filters\control_dependency\the 2018 ACS 5-year Detailed Tables product must be used instead of the 2018 ACS 1-year product\county coverage must be established under one consistent product-and-geography setup rather than by piecemeal fixes within the current table\all downstream county comparisons must use the same 5-year statewide baseline and the same all-counties geography\freeze\historical_window\2014-2018 American Community Survey 5-year estimates for North Carolina counties, using the North Carolina statewide row from the same release as the comparison baseline\answer_type\multi-row ordered table\,:inclusion_conditions\Use only North Carolina county data from the 2014-2018 American Community Survey 5-year estimates, and use the North Carolina statewide values from the same release as the comparison baseline.\Retain only counties with at least 4,000 households lacking internet access, and where the share of households without internet access, the share of households with a householder aged 65+ living alone, and the share of households without a vehicle are all higher than the statewide shares, and the median household income is lower than the statewide median.\Output the final results sorted alphabetically by county name, retaining the six required fields.\exclusion_conditions\Exclude results that incorrectly use the 2018 ACS 1-Year Estimates Detailed Tables or any other non-2014-2018 5-year release.\Exclude results that do not use the North Carolina statewide row from the same release as the comparison baseline.\Exclude any county that does not meet any one of the thresholds or statewide comparison conditions.\normalization\field_separator\|\record_separator\,\schema\County\NoInternetHH\NoInternetRatePct\SeniorLivingAloneSharePct\NoVehicleRatePct\MedianIncome\county\use the county display name including the word `County` and excluding a trailing state suffix\dedup_key\county FIPS / GEOID\percentage_fields\all percentage fields are plain numbers with exactly three decimal places and no percent sign\numeric_fields\household counts and median income are emitted as plain integers without commas\sorting_or_selection\county name alphabetical ascending\stop_condition\the North Carolina statewide threshold is fixed once from the designated table\all 100 North Carolina counties are covered under the same table/product settings\every surviving county is validated against all four required tables before emission\,:} |
| {:,:,:,:5,:,:,:,:,:State-Gated Retrieval\Use only the 2015-2019 American Community Survey 5-year estimates for New Mexico counties, and use the New Mexico statewide values from the same release as the comparison baseline.\Candidate counties must have at least 2,500 renter households and at least 500 limited-English-proficient households. Additionally, the LEP share, no-internet share, and no-vehicle share must all be higher than the statewide shares, and the median household income must be lower than the statewide median.\LEPHH must be computed by summing all four limited-English language groups from C16002. Finally, retain only counties where at least one of Severe Rent Burden or Crowded Renter is higher than the statewide value, and write StressPath according to the rule.\dependency_type\Data + Control\intra_chain\inter_chain\data_dependency\the 2015-2019 ACS 5-year county geography yields New Mexico county candidates and the statewide baseline row\C16002 and the downstream housing/income tables yield LEPHH, LEP share, no-internet share, no-vehicle share, severe rent burden, crowded-renter share, and median income for counties and the state baseline\final rows are produced only after the LEPHH numerator is computed from all four limited-English language groups and the downstream stress-path filters are applied\control_dependency\the LEPHH numerator must be computed from all four limited-English language groups in C16002 rather than from Spanish-only counts\county eligibility and StressPath labels must be determined from the LEPHH measure and statewide LEP baseline defined under that full four-group calculation\the RENT / CROWD / BOTH classification must be determined by evaluating the same county set against both downstream housing-stress branches\freeze\historical_window\2015-2019 American Community Survey 5-year estimates for New Mexico counties, using the New Mexico statewide row from the same release as the comparison baseline\answer_type\multi-row ordered table\,:inclusion_conditions\Use only the 2015-2019 American Community Survey 5-year estimates for New Mexico counties, and use the New Mexico statewide values from the same release as the comparison baseline.\Candidate counties must have at least 2,500 renter households and at least 500 limited-English-proficient households. Additionally, the LEP share, no-internet share, and no-vehicle share must all be higher than the statewide shares, and the median household income must be lower than the statewide median.\LEPHH must be computed by summing all four limited-English language groups from C16002. Finally, retain only counties where at least one of Severe Rent Burden or Crowded Renter is higher than the statewide value, and write StressPath according to the rule.\exclusion_conditions\Exclude results that treat Spanish-only limited-English households as the entire LEPHH count.\Exclude results that do not use the statewide baseline from the same release or that do not recalculate LEPHH/LEPSharePct/LEPShare_state.\Exclude counties that do not meet the final stress-path condition of RENT, CROWD, or BOTH.\normalization\field_separator\|\record_separator\,\schema\County\RenterHH\LEPHH\LEPSharePct\NoInternetPct\NoVehiclePct\SevereRentBurdenPct\CrowdedRenterPct\MedianIncome\StressPath\county\use the county display name including the word `County` and excluding a trailing state suffix\dedup_key\county FIPS / GEOID\LEPHH\sum the four designated limited-English-proficient renter-household cells before emitting the integer total\percentage_fields\all percentage fields are plain numbers with exactly three decimal places and no percent sign\numeric_fields\household counts and median income are emitted as plain integers without commas\StressPath\allowed_values\RENT\CROWD\BOTH\rule\emit BOTH only when both the severe-rent-burden and crowded-renter signals survive; otherwise emit the single surviving label\sorting_or_selection\county name alphabetical ascending\stop_condition\all candidate counties are validated under the same geography and product settings\the renter scope and LEP aggregation are checked before the final line is emitted\,:} |
| {:,:,:,:4,:not well\not at all\,:,:,:,:State-Gated Retrieval\Use only California county data from the 2013–2017 American Community Survey 5-year estimates, and use the California statewide value from the same release as the comparison baseline.\Retain only counties with a population aged 5–17 of at least 75,000, a SchoolAgeLEP of at least 2,000 with a share exceeding the statewide share, a no-internet share exceeding the statewide share, and a median household income below the statewide median.\SchoolAgeLEP must be the sum of the eight \\\ and \\\ cells across the four major language groups in B16004. The final county must also have at least two of the three offline support signals (Mobility, Benefits, Crowd) exceeding the statewide share.\dependency_type\Data + Control\intra_chain\inter_chain\data_dependency\The 2013–2017 ACS 5-year county geography yields California county candidates and the statewide baseline row.\B16004 and the downstream access/income/support tables yield SchoolAgeLEP, no-internet share, no-vehicle share, child-benefit share, crowded-renter share, and median income for counties and the statewide baseline.\Final rows are produced only after the SchoolAgeLEP numerator is computed from all eight 1980-target language cells and the downstream support-path branches are applied.\control_dependency\The SchoolAgeLEP numerator must be computed from the full eight-cell language set in B16004 rather than from only two Spanish LEP cells.\County eligibility and SupportPath labels must be determined from the LEP numerator and candidate set defined under that full eight-cell calculation.\The BENEFITS / CROWD / MOBILITY path labels must be determined by applying all three downstream support branches to the same county set.\freeze\historical_window\2013–2017 American Community Survey 5-year estimates for California counties, using the California statewide row from the same release as the comparison baseline.\answer_type\multi-row ordered table\,:inclusion_conditions\Use only California county data from the 2013–2017 American Community Survey 5-year estimates, and use the California statewide value from the same release as the comparison baseline.\Retain only counties with a population aged 5–17 of at least 75,000, a SchoolAgeLEP of at least 2,000 with a share exceeding the statewide share, a no-internet share exceeding the statewide share, and a median household income below the statewide median.\SchoolAgeLEP must be the sum of the eight \\\ and \\\ cells across the four major language groups in B16004. The final county must also have at least two of the three offline support signals (Mobility, Benefits, Crowd) exceeding the statewide share.\exclusion_conditions\Exclude results that only take the two Spanish LEP rows and do not combine all eight target cells.\Exclude results that do not re-run the B28011, B19013, B08201, B09010, and B25014 branches.\Exclude counties whose final SupportPath does not have at least two signals exceeding the statewide share.\normalization\field_separator\|\record_separator\,\schema\County\SchoolAgePop\SchoolAgeLEP\LEPSharePct\NoInternetPct\NoVehiclePct\ChildBenefitSharePct\CrowdedRenterPct\MedianIncome\SupportPath\county\Use the county display name including the word `County` and excluding a trailing state suffix.\dedup_key\county FIPS / GEOID\SchoolAgeLEP\Sum the eight designated school-age LEP cells before emitting the integer total.\percentage_fields\All percentage fields are plain numbers with exactly three decimal places and no percent sign.\numeric_fields\Population counts and median income are emitted as plain integers without commas.\SupportPath\Use only BENEFITS, CROWD, and MOBILITY, and when multiple labels apply join them with `+` in the fixed order BENEFITS+CROWD+MOBILITY.\sorting_or_selection\County name alphabetical ascending.\stop_condition\All candidate counties are validated under the same geography and product settings.\The school-age LEP aggregation and all three support-signal gates are checked before the final line is emitted.\,:} |
| {:,:,:,:7,:,:,:,:,:State-Gated Retrieval\Use only Pennsylvania county-level data from the 2012-2016 American Community Survey 5-year estimates, and use Pennsylvania statewide values from the same release as the comparison baseline.\Retain only counties with a population aged 65 and over of at least 15,000, at least 4,000 people aged 65 and over with ambulatory difficulty and a share higher than the statewide share, and a median household income below the statewide median.\The 65+ ambulatory difficulty numerator must combine the four target cells from B18105 for ages 65-74 and 75+, male and female. The final county must also have at least two of the four barriers—Alone, No Vehicle, Renter No Vehicle, Mobile Home—higher than the statewide values.\dependency_type\Data + Control\intra_chain\inter_chain\data_dependency\ACS tables B01001 and B18105 yield county senior-population totals and the full 65+ ambulatory-difficulty numerator.\Statewide comparison tables yield Pennsylvania baselines plus county barrier metrics for living alone, no vehicle, renter no vehicle, mobile home, and median income.\Final rows are produced only after the county candidate set is defined from the full ambulatory-difficulty numerator and the downstream six-table filters are applied.\control_dependency\The ambulatory-difficulty numerator must include the full 65+ population across 65-74 and 75+ and male and female cells, rather than only the 75+ cells.\Downstream county eligibility and barrier labels must be determined from the county candidate set defined under that full ambulatory numerator.\The at-least-two-barriers test must compare every county metric against the same Pennsylvania statewide baselines from the same ACS release.\freeze\historical_window\2012-2016 American Community Survey 5-year estimates for Pennsylvania counties, with Pennsylvania statewide baselines taken from the same release.\answer_type\multi-row ordered table\,:inclusion_conditions\Use only Pennsylvania county-level data from the 2012-2016 American Community Survey 5-year estimates, and use Pennsylvania statewide values from the same release as the comparison baseline.\Retain only counties with a population aged 65 and over of at least 15,000, at least 4,000 people aged 65 and over with ambulatory difficulty and a share higher than the statewide share, and a median household income below the statewide median.\The 65+ ambulatory difficulty numerator must combine the four target cells from B18105 for ages 65-74 and 75+, male and female. The final county must also have at least two of the four barriers—Alone, No Vehicle, Renter No Vehicle, Mobile Home—higher than the statewide values.\exclusion_conditions\Exclude results that take only the 75 years and over ambulatory difficulty cells and mistakenly treat them as the full 65+ value.\Exclude results that do not use the Pennsylvania statewide baselines from the same release.\Exclude counties whose final BarrierPath does not satisfy the condition of at least two barriers higher than the statewide values.\normalization\field_separator\|\record_separator\,\schema\County\SeniorPop\AmbulatoryDifficulty65Plus\AmbDiffPct\SeniorLivingAloneSharePct\NoVehiclePct\RenterNoVehiclePct\MobileHomePct\MedianIncome\BarrierPath\county\Use the county display name including the word `County` and excluding a trailing state suffix.\dedup_key\county FIPS / GEOID\percentage_fields\All percentage fields are plain numbers with exactly three decimal places and no percent sign.\numeric_fields\Population counts and median income are emitted as plain integers without commas.\BarrierPath\Use only ALONE, MOBILE_HOME, NO_VEHICLE, and RENTER_NO_VEHICLE. When multiple labels apply, join them with `+` in the fixed order ALONE+MOBILE_HOME+NO_VEHICLE+RENTER_NO_VEHICLE.\sorting_or_selection\county name alphabetical ascending\stop_condition\All candidate counties are validated under the same geography and product settings.\The senior ambulatory-difficulty gate and all barrier-signal gates are checked before the final line is emitted.\,:} |
| {:,:,:,:3,:,:,:,:,:State-Gated Retrieval\Screen only within the four South Texas divisions defined on the Texas divisional page: South Central, Upper Coast, South, Lower Valley.\Retain divisions that simultaneously satisfy: 1966-02 10-month Palmer Z-Index anomaly >= -0.05; 1966-12 24-month anomaly >= -0.10; 1966-03 1-month Heating Degree Days Value < 200°Df.\Sort by sum of the two Z anomalies descending, then by 1966-03 HDD Value ascending, then by division number ascending.\dependency_type\Data\intra_chain\inter_chain\data_dependency\The official Texas climate-division reference fixes the four South Texas division IDs and names that must be tracked across all later pages.\The 1966-02 10-month Palmer Z-Index anomaly page and the 1966-12 24-month Palmer Z-Index anomaly page jointly identify the four target South Texas divisions.\The 1966-03 1-month Heating Degree Days Value page supplies the last filter; final rows are produced only after intersecting the three divisional screens and then sorting by summed Z-anomaly score, HDD value, and division number.\freeze\historical_window\Texas climate divisions South Central, Upper Coast, South, and Lower Valley; metrics are fixed to the 1966-02 10-month Palmer Z-Index anomaly, 1966-12 24-month anomaly, and 1966-03 1-month Heating Degree Days Value pages.\answer_type\multi-row ordered table\,:inclusion_conditions\Screen only within the four South Texas divisions defined on the Texas divisional page: South Central, Upper Coast, South, Lower Valley.\Retain divisions that simultaneously satisfy: 1966-02 10-month Palmer Z-Index anomaly >= -0.05; 1966-12 24-month anomaly >= -0.10; 1966-03 1-month Heating Degree Days Value < 200°Df.\Sort by sum of the two Z anomalies descending, then by 1966-03 HDD Value ascending, then by division number ascending.\exclusion_conditions\Exclude any Texas divisional results outside the four specified South Texas divisions.\Exclude results that misread the 1966-03 HDD page filter field as Anomaly instead of Value.\Exclude any division that fails any time window or threshold condition.\normalization\field_separator\|\record_separator\,\schema\texas_division_no\division_name\feb_1966_10mo_z_text\dec_1966_24mo_z_text\mar_1966_hdd_text\moisture_score\texas_division_no\Canonical output uses the short Texas climate-division number; the scorer may accept an equivalent four-digit form, but the emitted oracle uses the short number.\raw_text_fields\The two Z-index anomaly fields and the March 1966 HDD field preserve the raw displayed text; comparison may ignore optional trailing HDD unit text such as `°Df`.\moisture_score\Compute as February 1966 10-month Z-index anomaly plus December 1966 24-month Z-index anomaly, formatted with exactly two decimal places.\sorting_or_selection\primary\moisture_score descending\secondary\March 1966 HDD actual value ascending after numeric normalization\tertiary\Texas division number ascending\eligibility_rule\Each Texas division is counted once per requested window; if any required window value is missing, exclude the division.\stop_condition\All candidate Texas divisions are checked under the same base period and anomaly windows.\The unique ordered output string is resolved before emission.\,:} |
| {:,:,:,:2,:rebound score\<Texas division number>|<4-digit division ID>|<division name>|<1907-04 10-month Z-Index anomaly raw text>|<1923-04 10-month Z-Index anomaly raw text>|<1915-05 2-month HDD actual value raw text>|<1922-10 5-month HDD actual value raw text>|<rebound score with 2 decimal places>\,:,:<Texas division number>|<4-digit division ID>|<division name>|<1907-04 10-month Z-Index anomaly raw text>|<1923-04 10-month Z-Index anomaly raw text>|<1915-05 2-month HDD actual value raw text>|<1922-10 5-month HDD actual value raw text>|<rebound score with 2 decimal places>\,:,:State-Gated Retrieval\Retain only Texas divisions that satisfy all four historical conditions: the 10-month Palmer Z-Index anomaly ending April 1907 is no higher than -0.30; the 10-month anomaly ending April 1923 is no lower than -0.20; the 2-month HDD Value ending May 1915 is less than 50°Df; the 5-month HDD Value ending October 1922 is less than 20°Df.\Rebound score is defined as the April 1923 10-month anomaly minus the April 1907 10-month anomaly.\Sort by rebound score descending, then by October 1922 5-month HDD Value ascending, then by Texas division number ascending.\dependency_type\Data + Control\intra_chain\inter_chain\data_dependency\Texas divisional pages yield the 1907-04 and 1923-04 10-month Palmer Z-Index anomalies plus the 1915-05 2-month and 1922-10 5-month HDD values for each division.\The rebound score and the two shoulder-season HDD values are computed only after the same set of candidate divisions survives all four historical filters.\The final output is produced after sorting the surviving divisions by rebound score, then 1922-10 HDD value, then division number.\control_dependency\The workflow must cover the 1907, 1923, and 1915 pages as well as the 1922-10 5-month HDD page.\Both HDD pages must be read from Value rather than Anomaly.\The final rebound ranking must be determined after the 1922-10 5-month HDD page excludes 7 South Central and 8 Upper Coast, using ReboundScore as the ordering basis.\freeze\historical_window\Texas climate divisions evaluated on the 1907-04 and 1923-04 10-month Palmer Z-Index anomaly pages, the 1915-05 2-month HDD Value page, and the 1922-10 5-month HDD Value page.\answer_type\mixed multi-line output\,:inclusion_conditions\Retain only Texas divisions that satisfy all four historical conditions: the 10-month Palmer Z-Index anomaly ending April 1907 is no higher than -0.30; the 10-month anomaly ending April 1923 is no lower than -0.20; the 2-month HDD Value ending May 1915 is less than 50°Df; the 5-month HDD Value ending October 1922 is less than 20°Df.\Rebound score is defined as the April 1923 10-month anomaly minus the April 1907 10-month anomaly.\Sort by rebound score descending, then by October 1922 5-month HDD Value ascending, then by Texas division number ascending.\exclusion_conditions\Exclude results that stop after filtering only the 1907, 1923, and 1915 pages without also reading the 1922-10 5-month HDD page.\Exclude results that incorrectly read the 1915 or 1922 HDD page as Anomaly instead of Value.\Exclude any division that does not satisfy all historical window conditions.\normalization\field_separator\|\record_separator\,\schema\texas_division_no\division_id_4digit\division_name\apr_1907_10mo_z_text\apr_1923_10mo_z_text\may_1915_2mo_hdd_text\oct_1922_5mo_hdd_text\rebound_score\texas_division_no\Emit the short Texas climate-division number in field 1 and the four-digit division ID in field 2.\raw_text_fields\The two Z-index anomaly fields preserve the raw displayed text; the two HDD fields preserve the raw displayed text, but comparison may ignore an optional trailing `°Df` suffix.\rebound_score\Compute as the April 1923 10-month Z-index anomaly minus the April 1907 10-month Z-index anomaly, formatted with exactly two decimal places.\sorting_or_selection\primary\rebound_score descending\secondary\October 1922 HDD actual value ascending after numeric normalization\tertiary\Texas division number ascending\answer_set_note\The current oracle stores two equivalent answer sets: one preserves the HDD unit suffixes and one strips them; treat the two HDD columns as equivalent after optional `°Df` removal.\stop_condition\All candidate Texas divisions are checked under the same anomaly and HDD windows.\The unique ordered output string is resolved before emission.\,:} |
| {:,:,:,:5,:,:,:,:,:State-Gated Retrieval\Screen only the five substances specified in the task, and each target substance page must explicitly contain the record “Arts and crafts/Office supplies - body paint - tattoo ink, detected, Europe.”\Check whether each of the five specified dated keywords appears, and record the earliest and latest year-month among the hits.\Then independently select one daily-consumer route and one production-material route, the former following Formulation → Article → Occupation and the latter following Occupation → Article → Formulation, and refine each route separately to its deepest level.\dependency_type\Data + Control\intra_chain\inter_chain\data_dependency\each chemical page yields the five dated historical keywords, the Europe body-paint/tattoo-ink qualifying record, and the earliest/latest dated-hit months\PUC category paths yield one consumer-side route chosen by Formulation -> Article -> Occupation and one source-side route chosen independently by Occupation -> Article -> Formulation\final rows are produced only after the consumer-side and source-side routes are selected and refined separately for each chemical\control_dependency\the source-side route must be selected independently from the consumer-side route\for each chemical, the second route refinement must follow the production-material priority order rather than the daily-consumer route\family/type/general outputs must be determined only after both route selections are made independently\freeze\historical_window\current ChemExpo chemical pages and PUC category paths for the five named substances, using the dated keyword states through 2023-01 and the current Europe body-paint/tattoo-ink record\answer_type\multi-row ordered table\,:inclusion_conditions\Screen only the five substances specified in the task, and each target substance page must explicitly contain the record “Arts and crafts/Office supplies - body paint - tattoo ink, detected, Europe.”\Check whether each of the five specified dated keywords appears, and record the earliest and latest year-month among the hits.\Then independently select one daily-consumer route and one production-material route, the former following Formulation → Article → Occupation and the latter following Occupation → Article → Formulation, and refine each route separately to its deepest level.\exclusion_conditions\Exclude substances that do not contain the specified Europe body-paint/tattoo-ink record.\Exclude results that directly reuse the same daily-consumer route for the production-material side.\Exclude results that do not independently select and refine both routes according to the specified priority orders.\normalization\field_separator\|\record_separator\,\schema\DTXSID\hit_count\earliest\latest\consumer_kind\consumer_general_puc_id\consumer_terminal_puc_id\consumer_level\consumer_products\material_kind\material_general_puc_id\material_terminal_puc_id\material_level\material_products\dated_hits\count only the exact dated keyword strings defined in the source rubric\date_format\YYYY-MM\kind\the kind fields use the canonical labels Article, Formulation, or Occupation\level\the level fields use the canonical labels general, family, or type\products_source\the Products counts come from the local chemical-detail node for the selected branch, not from an overall aggregate shown elsewhere on the site\sorting_or_selection\primary\hit_count descending\secondary\consumer_products descending\tertiary\DTXSID ascending\,:} |
| {:,:,:,:6,:Toys and children's products\"; any keyword containing \"artificial_saliva\" or \"artificial_sweat\". Next, check whether any of the two manufacturing-source-related label categories appear; score 1 point for each category that appears: drinking_water, Europe, manufacturing, plastic_additive; Europe, plastic_additive. Additionally, separately count how many of the following four dated keywords appear: Indirect additives food contact (10/2018); OEHHA Proposition 65 (3/2019); WA Children's Safe Product Act (4/2020); OEHHA Proposition 65 (1/2023). Record the number of hits and the earliest and latest year-month among the hits. Then determine the direction based on which side has a higher score: if the material-contact score is strictly greater than the manufacturing-source score, search for the most representative category for this substance in the order Article → Formulation → Occupation; otherwise, search in the order Occupation → Article → Formulation. After determining the category, from this level onward, at each level only follow the subcategory with the highest product count, continuing until no further subdivision is possible; if there is a tie at the same level, choose the one with the smaller PUC ID.start_urlhttps://comptox.epa.gov/chemexpo/output_formatFinally, sort all substances by dated hit count descending; for ties, sort by product count of the selected node descending. Output format: <DTXSID>|<contactHits>|<sourceHits>|<datedHits>|<earliest>|<latest>|<prioritySide>|<PUCkind>|<generalPUCID>|<terminalPUCID>|<level>|<Products>|<PUCDocuments>|<PUCCuratedChemicals>. Separate results with commas. If datedHits is 0, write NA for both earliest and latest.oracle_answerDTXSID5020607|5|2|4|2018-10|2023-01|contact|Article|309|493|family|1|23|58,DTXSID5020601|4|1|3|2018-10|2023-01|contact|Article|309|309|general|1|5836|1827,DTXSID6020802|4|2|1|2018-10|2018-10|contact|Article|313|313|general|15|1104|675,DTXSID8021484|3|1|1|2018-10|2018-10|contact|Occupation|321|321|general|8|28313|4200,DTXSID0021414|5|0|1|2018-10|2018-10|contact|Article|309|309|general|3|5836|1827,DTXSID20872645|0|1|0|NA|NA|source|Occupation|318|381|family|5|274|107metadata{\:[\,\,\],\:\,\:true,\:true,\:[\,\,\],\:[\,\,\],\:{\:\},\:\}rubric{\:[\,\,\],\:[\,\,\],\:{\:\,\:\,\:[\,\,\,\,\,\,\,\,\,\,\,\,\,\],\:\,\:\,\:\,\:\,\:\,\:\,\:\,\:{\:\,\:\,\:\}}}all_involved_urlsnull |
| task_idcomptox_003domainCOMPTOX_CHEMEXPOautonomy_typeordered tableoracle_output_cardinalityinstructionIn ChemExpo, first select Formulation categories with a total product count of 15,000 or more. Then, within each category, following the website's default display order, identify the first product family that meets these criteria: the family itself has no assumed attributes and its total product count is at least 3,000; additionally, the family must have at least two distinct specific product types, one with allowed attributes including aerosol and the other including child, and both types must have empty assumed attributes. Next, within each of these two specific product types, compute the product share (type product count divided by family cumulative product count). If shares are equal, compare their Curated Chemicals counts and select the higher one; if still equal, select the type with the smaller PUC ID. After completing the selection on both sides, retain only the better of the two candidate types: first prioritize the type with the higher product share; if shares are equal, compare Product Count (higher wins); if still equal, compare Curated Chemicals (higher wins); if still equal, select the type with the smaller PUC ID.","start_url":"https://comptox.epa.gov/chemexpo/visualizations/","output_format":"Output results sorted by final share in descending order, formatted as: <generalPUCID>|<familyPUCID>|<typePUCID>|<side>|<share>|<CuratedChemicals>. Join records with commas and no spaces. If no qualifying items exist, output NONE.","oracle_answer":"48|42|62|child|2160/3531|583,137|197|210|child|6934/16904|1135","metadata":"{\"State-Gated Retrieval\":[\"First, retain only Formulation categories with a total product count of 15,000 or more.\",\"Within each candidate family, both an aerosol-side type and a child-side type must be found: the family itself must have no assumed attributes and a total product count of at least 3,000; both types must also have empty assumed attributes.\",\"All final selections within a family and between the two sides must follow the specified tie-break rules: share, Product Count, Curated Chemicals, and PUC ID.\"],\"dependency_type\":\"Data + Control\",\"intra_chain\":true,\"inter_chain\":true,\"data_dependency\":[\"The Formulation summary yields the high-product candidate families whose total product count is at least 15,000\",\"Within each family, lower-level type pages yield the aerosol and child candidates, their assumed/allowed attributes, product counts, shares, and curated-chemical counts\",\"Final rows are produced only after a family survives both aerosol and child branches and the winning type is selected by the layered share/product/chemical/PUC tie-break rules\"],\"control_dependency\":[\"Both the aerosol and child branches within the same family must be fully checked before that family can be retained\",\"If one side is missing, the workflow must reject the family and continue to the next sibling family rather than stop at the first partial hit\",\"The final winner can be selected only after both sides have been independently optimized within the family and then compared against each other\"],\"freeze\":{\"historical_window\":\"Current ChemExpo Formulation families with total product count >= 15,000, evaluated on family/type attributes and current product/chemical counts\"},\"answer_type\":\"multi-row ordered table\"}","rubric":"{\"inclusion_conditions\":[\"First, retain only Formulation categories with a total product count of 15,000 or more.\",\"Within each candidate family, both an aerosol-side type and a child-side type must be found: the family itself must have no assumed attributes and a total product count of at least 3,000; both types must also have empty assumed attributes.\",\"All final selections within a family and between the two sides must follow the specified tie-break rules: share, Product Count, Curated Chemicals, and PUC ID.\"],\"exclusion_conditions\":[\"Exclude results that stop as soon as a usable type is found on either side without requiring the same family to cover both aerosol and child sides.\",\"Exclude results that stop as soon as one side is missing instead of continuing to check sibling families at the same level.\",\"Exclude results that do not follow the specified layered tie-break rules to select the winning type within a family and overall.\"],\"normalization\":{\"field_separator\":\"|\",\"record_separator\":\",\",\"schema\":[\"generalPUCID\",\"familyPUCID\",\"typePUCID\",\"side\",\"share\",\"CuratedChemicals\"],\"dedup_key\":\"(generalPUCID, familyPUCID, typePUCID, side)\",\"side\":\"Emit the lower-case branch label exactly as determined in the source workflow\",\"share\":\"Canonical output uses the raw fraction `type Products / family Cumulative Products` rather than a decimal\",\"CuratedChemicals\":\"Emit the integer count from the selected type-level node\",\"sorting_or_selection\":{\"primary\":\"Share descending by numeric fraction value\",\"secondary\":\"Type-level Products descending\",\"tertiary\":\"CuratedChemicals descending, then PUC IDs ascending as tie-breaks\"}}}","all_involved_urls":"null"} |
| {"task_id":"comptox_004","domain":"COMPTOX_CHEMEXPO","autonomy_type":"ordered table","oracle_output_cardinality":5,"instruction":"In ChemExpo, for the five substances Di(2-Ethylhexyl) Phthalate, Melamine, Phthalic Acid, Tris(2-Ethylhexyl) Phosphate, and Triphenyl Phosphates Tert-Butylated, check whether their historical labels align with their current uses. Follow the steps below. First, restrict the scope to these five substances. For each substance, collect three sets of baseline assessment data. The first set is the \"material contact\" score: check for the presence of the following five label categories, scoring 1 point for each category hit: Indirect additives food contact (10/2018); Europe, Food contact items; Food contact items, Japan; any keyword containing \"Toys and children's products\artificial_saliva\artificial_sweat\production source\drinking_water, Europe, manufacturing, plastic_additive\Europe, plastic_additive\,:,:,:,:State-Gated Retrieval\Operate only on the five specified substances, and first compute materialHits, sourceHits, and datedHits from the keyword-set block.\The historically recommended category must be derived independently from these historical signals; the current dominant category must be derived separately from the current use path.\Finally, compare only whether the historically recommended category and the current dominant category align at the coarse side level, and mark ALIGNED or MISALIGNED accordingly.\dependency_type\Data + Control\intra_chain\inter_chain\data_dependency\each chemical page yields materialHits, sourceHits, datedHits, and the historical recommendation side implied by the keyword set\an independent current PUC category path yields the current dominant category, path labels, and classification counts for the same chemical\final rows are produced only after the historical recommendation and current dominant category are computed independently and then compared at the coarse side level\control_dependency\the historical recommendation must be determined from the keyword-set evidence rather than from the current largest category\historical recommendation must be computed from the keyword-set evidence before alignment is judged\ALIGNED / MISALIGNED status is valid only after the historical and current category directions have been derived independently and then compared\freeze\historical_window\current ChemExpo chemical pages and PUC category paths for the five named substances, using the specified historical keyword-set evidence through 2023-01\answer_type\multi-row ordered table\,:inclusion_conditions\Operate only on the five specified substances, and first compute materialHits, sourceHits, and datedHits from the keyword-set block.\The historically recommended category must be derived independently from these historical signals; the current dominant category must be derived separately from the current use path.\Finally, compare only whether the historically recommended category and the current dominant category align at the coarse side level, and mark ALIGNED or MISALIGNED accordingly.\exclusion_conditions\Exclude results that simply take the current largest category and treat it as the historically recommended category.\Exclude results that default to ALIGNED without independently deriving the historically recommended category.\Exclude results that fail to compute the historical and current paths separately, instead mixing signals in a single path.\normalization\field_separator\|\record_separator\,\schema\DTXSID\materialHits\sourceHits\datedHits\earliest\latest\historySide\recommendedKind\recommendedGeneral\recommendedTerminal\recommendedLevel\recommendedProducts\currentKind\currentGeneral\currentTerminal\currentLevel\currentProducts\status\dated_hits\count only the exact dated keyword strings defined in the source rubric\date_format\YYYY-MM; if datedHits is 0, earliest and latest are both `NA`\historySide\emit the lower-case historical winner `material` or `source`\recommended_fields\these fields describe the historically suggested path\current_fields\these fields describe the current dominant path\kind\the kind fields use the canonical labels Article, Formulation, or Occupation\level\the level fields use the canonical labels general, family, or type\status\emit `ALIGNED` or `MISALIGNED` exactly\sorting_or_selection\primary\status with MISALIGNED before ALIGNED\secondary\datedHits descending\tertiary\currentProducts descending\quaternary\DTXSID ascending\,:} |
| {:,:,:,:3,:,:,:,:,:State-Gated Retrieval\Only consider administrative enforcement actions on the CFPB Enforcement actions page whose Initial filing date falls between 2022-01-01 and 2024-12-31.\For each of the three product categories (Deposits, Furnishing, Prepaid), search in the current default order from top to bottom for the first action that simultaneously satisfies: the page provides a Consent Order and a Press release; the Products label corresponds one-to-one with the target category; the action detail page explicitly states consumer redress > $5,000,000 and also states a civil money penalty.\Finally, extract the first-occurrence amounts and page numbers of consumer redress and civil money penalty from the correct Consent Order.\dependency_type\Data + Control\intra_chain\inter_chain\data_dependency\Enforcement-action list pages yield product-branch candidate actions in current default order.\Each action detail page and linked consent-order PDF yield product singularity, document-stack presence, redress/civil-money-penalty amounts, and first-occurrence pages.\Final rows are emitted only after identifying, for each product, the first action that survives detail-page and PDF validation.\control_dependency\List-level \\\ hits must be validated on the detail page for product singularity and money thresholds.\Prepaid candidates force document-level disambiguation among Consent Order, Stipulation, and terminating-order PDFs.\Per-product first-match selection must explicitly skip rejected candidates and validate downstream PDFs before it is finalized.\freeze\historical_window\Administrative enforcement actions with Initial filing date from 2022-01-01 through 2024-12-31; current default case ordering and currently linked action documents.\answer_type\multi-row ordered table\,:inclusion_conditions\Only consider administrative enforcement actions on the CFPB Enforcement actions page whose Initial filing date falls between 2022-01-01 and 2024-12-31.\For each of the three product categories (Deposits, Furnishing, Prepaid), search in the current default order from top to bottom for the first action that simultaneously satisfies: the page provides a Consent Order and a Press release; the Products label corresponds one-to-one with the target category; the action detail page explicitly states consumer redress > $5,000,000 and also states a civil money penalty.\Finally, extract the first-occurrence amounts and page numbers of consumer redress and civil money penalty from the correct Consent Order.\exclusion_conditions\Exclude actions where the Products label is not a single target category, or where the action appears relevant only at the list level but the detail page does not confirm it.\Exclude actions that do not clearly satisfy the redress > $5M and independent civil money penalty conditions.\Exclude cases where a Stipulation or Order Terminating the Consent Order is mistakenly used in place of the required Consent Order.\normalization\field_separator\|\record_separator\,\schema\Product\Respondent\Docket\FilingDate\RedressUSD\RedressPage\PenaltyUSD\PenaltyPage\product_order\Alphabetical by Product name\fallback_row\If a product has no qualifying case, emit `<Product>|NONE`\date_format\FilingDate uses YYYY-MM-DD\money_fields\Canonical output preserves the dollar sign and commas from the oracle, but equivalent comparison may ignore `$` and commas and compare the numeric USD amount.\page_fields\Page fields use the canonical `pN` form.\document_scope\Extract both money amounts from the Consent Order PDF only; do not substitute a stipulation, termination order, or other companion document.\selection_rule\For each product, keep the first action in the default filtered order that fully satisfies every gate in the source rubric.\,:} |
| {:,:,:,:2,:belongs to the same survey series.\,:,:,:,:State-Gated Retrieval\First, locate the CFPB household financial well-being annual report series whose latest version links directly to 2023 and 2022.\Using the 2022 edition as the baseline, retain only topics that appear in 2022 but are no longer retained in subsequent annual reports, and exclude the topic specifically marked as \\\\For each topic, find the earliest non-series pivot report from 2022 or earlier, then count its directly listed related resources, how many link back to the same survey series, how many qualify under the appendix rule, and the appendix span of the final qualifying report.\dependency_type\Data + Control\intra_chain\inter_chain\data_dependency\current CFPB Reports pages yield the 2024 -> 2023 -> 2022 annual-survey chain and the 2022 topic baseline\report detail pages and report-search results yield first-disappearance year, pivot report title/date, related-resource counts, and links back to the annual-survey series\final rows are produced only after separating the first jump target from later return-to-series counting\control_dependency\requiring the pivot report itself to link back to the same survey series incorrectly rules out valid topic rows\title extraction must use the pivot detail page rather than search-result snippets\related-resource expansion must continue even when the pivot report itself does not link back to the same survey series\freeze\historical_window\currently linked 2024, 2023, and 2022 CFPB household financial well-being annual reports; topic disappearance is judged over that three-year chain, and pivot-report search is restricted to 2022-and-earlier reports\answer_type\multi-row ordered table\,:inclusion_conditions\First, locate the CFPB household financial well-being annual report series whose latest version links directly to 2023 and 2022.\Using the 2022 edition as the baseline, retain only topics that appear in 2022 but are no longer retained in subsequent annual reports, and exclude the topic specifically marked as \\\\For each topic, find the earliest non-series pivot report from 2022 or earlier, then count its directly listed related resources, how many link back to the same survey series, how many qualify under the appendix rule, and the appendix span of the final qualifying report.\exclusion_conditions\Exclude topics that are still retained in subsequent annual reports.\Exclude results that incorrectly treat \\\ as a hard condition for column 3.\Exclude search results that directly belong to the same survey series or reports from after 2022.\normalization\field_separator\|\record_separator\,\schema\topic\first-missing-year\pivot-report-or-NONE\pivot-date-or-NONE\direct-related-count\same-series-count\qualified-count\representative-report-or-NONE\publication-date-or-NONE\appendix-span-or-NONE\topic_order\alphabetical by topic name\date_format\pivot-date and publication-date use YYYY-MM-DD when present\appendix_span\canonical span uses compact labels such as `A-D`\none_rule\if qualified-count is 0, representative-report, publication-date, and appendix-span are all `NONE`\,:} |
| {:,:,:,:2,:,:,:,:,:State-Gated Retrieval\Must process the two 2020 Debt Collection Practices (Regulation F) final rules separately, not merge them into one.\For each wave, separately confirm: issue date, corresponding proposal chain and public comment deadlines, whether the 2021 delayed-effective-date proposed rule covers it and its deadline, final effective date, corresponding implementation quick-reference PDF, and the first heading after the introduction in that PDF.\The detail page, proposal chain, and summary PDF for the October wave and the December wave must each be independently verified.\dependency_type\Data + Control\intra_chain\inter_chain\data_dependency\final-rules index pages and rule-development pages yield the two 2020 Regulation F final-rule waves and their proposal branches\each final-rule detail page, delayed-effective-date proposal page, and implementation quick-reference PDF yields the wave-specific dates, linked proposals, comment deadlines, and first summary heading\final rows are produced only after the October and December waves are traced separately end to end\control_dependency\two same-title final rules force correction from name-based collapsing to issue-date branching\proposal-chain validation must be done per wave; otherwise the later December chain contaminates the October row\implementation PDFs must be matched back to the correct wave rather than accepting the latest quick-reference page\freeze\historical_window\the 2020-10 and 2020-12 Debt Collection Rule final-rule waves, their linked 2020 proposal pages, the 2021 delayed-effective-date proposal, and the current implementation-page quick-reference PDFs\answer_type\multi-row ordered table\,:inclusion_conditions\Must process the two 2020 Debt Collection Practices (Regulation F) final rules separately, not merge them into one.\For each wave, separately confirm: issue date, corresponding proposal chain and public comment deadlines, whether the 2021 delayed-effective-date proposed rule covers it and its deadline, final effective date, corresponding implementation quick-reference PDF, and the first heading after the introduction in that PDF.\The detail page, proposal chain, and summary PDF for the October wave and the December wave must each be independently verified.\exclusion_conditions\Exclude results that merge the two identically named 2020 final rules into one.\Exclude results that project the December wave's proposal chain or executive-summary heading onto the October wave.\",\"Exclude results that do not independently verify each final-rule detail page and its corresponding PDF.\"],\"normalization\":{\"field_separator\":\"|\",\"record_separator\":\",\",\"schema\":[\"wave\",\"final-rule-date\",\"proposal-chain\",\"proposal-comment-closes\",\"delay-proposal-applied\",\"delay-proposal-comments-close\",\"controlling-effective-date\",\"quick-reference-pdf\",\"summary-heading\"],\"wave\":\"wave is the final-rule issue month in YYYY-MM\",\"date_format\":\"all standalone date fields use YYYY-MM-DD\",\"proposal_chain\":\"canonical chain joins proposal milestones with `+`, using pieces such as `2019-05-07 main` or `2020-02-21 supplemental` in chronological order\",\"proposal_comment_closes\":\"join multiple comment-close dates with `+` in the same order as proposal_chain\",\"delay_proposal_applied\":\"emit `YES` when a delay proposal is part of the chain; otherwise emit `NO`\",\"quick_reference_pdf\":\"use the displayed quick-reference PDF title, not the filename\",\"summary_heading\":\"use the first substantive heading from the final-rule PDF\",\"sorting_or_selection\":\"final-rule-date ascending\"}}","all_involved_urls":"null"} |
| {"task_id":"consumerfinance_004","domain":"CFPB_REPORTS","autonomy_type":"ordered table","oracle_output_cardinality":2,"instruction":"On the CFPB Enforcement Actions page, identify administrative enforcement actions filed in 2024 and subsequently closed, separately for the product categories Credit Cards and Deposits. Then extract the corresponding clause information linking the terminating order to the original Consent Order. Specifically, first restrict the search to administrative enforcement actions on the CFPB Enforcement Actions page whose initial filing date falls between 2024-01-01 and 2024-12-31. Then filter separately for the two product categories: Credit Cards and Deposits. For each category, retain only cases that simultaneously meet the following conditions: the case page shows the case has been closed, with a status of Expired, Terminated, or Vacated; the case page provides all four of these document types: Consent Order, Stipulation, Order Terminating the Consent Order, and Press Release; the page lists exactly one product category, and that single category is the one you are currently targeting. That is, if you are searching for Credit Cards, the page must list only that item and no other categories; the same requirement applies when searching for Deposits. After filtering, each product category may still have more than one case. For each category, retain only the case with the latest initial filing date; if multiple cases share the same filing date, select the one with the smaller docket number. Next, open the Order Terminating the Consent Order for the selected case and record the following three pieces of information: the termination date on this document; the passage in the document that states the legal authority for terminating the consent order, i.e., the termination authority, recording only the original text of its first occurrence; the PDF physical page number where this termination authority first appears. Also, within the text of this terminating order, find the first explicit reference to a paragraph number from the original Consent Order; if it reads \"Paragraphs X and Y,\" take only X. Then, return to the original Consent Order document for this case, locate the referenced paragraph, and record the section heading under which it falls.","start_url":"https://www.consumerfinance.gov/enforcement/actions/","output_format":"Finally, output a single-line string in alphabetical order by product name: <Product>|<Respondent>|<Docket>|<InitialFilingDate>|<TerminationDate>|<TerminationAuthority>|p<TermAuthorityPage>|<ReferencedParagraph>|<ConsentSectionHeading>. If a product has no qualifying case, output: <Product>|NONE","oracle_answer":"Credit Cards|Apple Inc.|2024-CFPB-0012|2024-10-23|2025-09-22|12 U.S.C. § 5563(b)(3)|p1|83|ADMINISTRATIVE PROVISIONS,Deposits|Navy Federal Credit Union|2024-CFPB-0014|2024-11-07|2025-07-01|12 U.S.C. § 5563(b)(3)|p1|106|ADMINISTRATIVE PROVISIONS","metadata":"{\"State-Gated Retrieval\":[\"Consider only administrative enforcement actions with an Initial filing date in 2024, and search separately in the Credit Cards and Deposits categories.\",\"Cases must have a status of Expired, Terminated, or Vacated, and the page must simultaneously provide a Consent Order, Stipulation, Order Terminating the Consent Order, and Press Release, and the Products field must correspond one-to-one with the target category.\",\"Finally, for each category retain only the most recent qualifying case determined by Initial filing date, and compile the corresponding clause information linking the terminating order to the original Consent Order.\"],\"dependency_type\":\"Data + Control\",\"intra_chain\":true,\"inter_chain\":true,\"data_dependency\":[\"enforcement-action list pages yield 2024 candidate actions by product branch\",\"each action detail page and linked Consent Order / Stipulation / Order Terminating the Consent Order / Press Release documents yield status, initial filing date, single-product validity, and clause references\",\"final rows are selected only after per-product comparison on Initial filing date and document-stack validation\"],\"control_dependency\":[\"\\\"latest\\\" must be determined by Initial filing date rather than termination date\",\"product filters on the list page are insufficient until the action detail page confirms a single matching product\",\"clause extraction in the terminating order must stay aligned with the original Consent Order rather than a superficially similar section\"],\"freeze\":{\"historical_window\":\"administrative enforcement actions with Initial filing date from 2024-01-01 through 2024-12-31; current action status and currently linked Consent Order / Stipulation / terminating-order / Press Release documents\"},\"answer_type\":\"multi-row ordered table\"}","rubric":"{\"inclusion_conditions\":[\"Consider only administrative enforcement actions with an Initial filing date in 2024, and search separately in the Credit Cards and Deposits categories.\",\"Cases must have a status of Expired, Terminated, or Vacated, and the page must simultaneously provide a Consent Order, Stipulation, Order Terminating the Consent Order, and Press Release, and the Products field must correspond one-to-one with the target category.\",\"Finally, for each category retain only the most recent qualifying case determined by Initial filing date, and compile the corresponding clause information linking the terminating order to the original Consent Order.\"],\"exclusion_conditions\":[\"Exclude results that use termination date instead of Initial filing date to determine \\\"latest.\\\"\",\"Exclude cases with an incomplete document stack or where Products is not a single target category.\",\"Exclude cases where the required clause correspondence between the terminating order and the original Consent Order cannot be established.\"],\"normalization\":{\"field_separator\":\"|\",\"record_separator\":\",\",\"schema\":[\"Product\",\"Respondent\",\"Docket\",\"InitialFilingDate\",\"TerminationDate\",\"TerminationAuthority\",\"TermAuthorityPage\",\"ReferencedParagraph\",\"ConsentSectionHeading\"],\"product_order\":\"alphabetical by Product name\",\"fallback_row\":\"if a product has no qualifying case, emit `<Product>|NONE`\",\"date_format\":\"InitialFilingDate and TerminationDate use YYYY-MM-DD\",\"termination_authority\":\"use the canonical statute string extracted from the termination order\",\"page_fields\":\"TermAuthorityPage uses the canonical `pN` form\",\"ReferencedParagraph\":\"emit the paragraph number as a plain numeric string\",\"ConsentSectionHeading\":\"emit the uppercase heading of the consent-order section containing the referenced paragraph\"}}","all_involved_urls":"null"} |
| {"task_id":"consumerfinance_005","domain":"CFPB_COMPLAINT_DATABASE","autonomy_type":"ordered table","oracle_output_cardinality":2,"instruction":"I am preparing an internal outreach memo on historical complaint clusters and need to track a specific set of complaints: from the CFPB Complaint Database, within the current product bucket covering credit reporting and other personal consumer reports, between 2018-01-01 and 2021-12-31, tagged as Servicemember, with a public narrative, and where the text explicitly mentions medical bills or doctors. Please provide two representative complaint clusters: the first is the peak month of this cohort; the second is the latest month in this cohort whose complaint count is still at least half of the peak. For each cluster, report the total complaints for that month, the state with the highest complaint count and its count, the company with the highest complaint count within that state and month and its count, and the earliest complaint received by that company in that state and month, including its ID, date, Issue, and Sub-issue. Output a single line with Peak first, then LatestHalfPeak, using the format: <Cluster>|<Month>|<MonthComplaints>|<TopState>|<StateComplaints>|<TopCompany>|<CompanyComplaints>|<EarliestComplaintID>|<EarliestDate>|<Issue>|<SubIssue>.","start_url":"https://www.consumerfinance.gov/data-research/consumer-complaints/","output_format":"Output a single line with Peak first, then LatestHalfPeak, using the format: <Cluster>|<Month>|<MonthComplaints>|<TopState>|<StateComplaints>|<TopCompany>|<CompanyComplaints>|<EarliestComplaintID>|<EarliestDate>|<Issue>|<SubIssue>. Join the Peak and LatestHalfPeak clusters with a comma. For ties: months are ordered chronologically; states are ordered alphabetically by state abbreviation; companies are ordered alphabetically by company name; for the earliest complaint, use the earliest date, then the smallest Complaint ID if dates are identical.","oracle_answer":"Peak|2018-02|46|CA|28|TRANSUNION INTERMEDIATE HOLDINGS, INC.|9|2806154|2018-02-07|Improper use of your report|Reporting company used your report improperly,LatestHalfPeak|2021-11|23|CA|12|EQUIFAX, INC.|3|4928768|2021-11-19|Problem with a credit reporting company's investigation into an existing problem|Their investigation did not fix an error on your reportmetadata{\:[\,\,\],\:\,\:true,\:true,\:[\,\,\],\:[\,\,\],\:{\:\},\:\}rubric{\:[\,\,\],\:[\,\,\],\:{\:\,\:\,\:[\,\,\,\,\,\,\,\,\,\,\],\:[\,\],\:{\:\,\:\},\:\,\:\,\:\,\:\}}all_involved_urlsnull |
| task_idconsumerfinance_006domainCFPB_COMPLAINT_DATABASEautonomy_typeordered tableoracle_output_cardinalityinstructionI am writing a crypto-complaint concentration memo. To verify whether complaint peaks within a fixed three-year window are driven only by a few top platforms, I need you to step through the CFPB complaint database and extract relevant data.\n\nFirst, define the underlying data scope. Using the current CFPB complaint form product taxonomy, determine which product bucket Virtual currency currently falls under. Once the bucket is confirmed, strictly limit the time range to 2019-08-26 through 2022-08-26 and retain only complaint records classified under the Virtual currency subcategory within that product bucket. This forms the initial cohort.\n\nSecond, identify two characteristic peak months within this cohort. Step 1: Aggregate complaints by month across the full cohort (with no companies removed) and identify the original peak month (AllCompanies). Step 2: Identify the top two companies by complaint count in that original peak month. Remove all complaint records of these two companies from the entire three-year cohort (not just from the single peak month). Then re-aggregate complaints by month on the remaining data to find the residual peak month (ResidualAfterTop2).\n\nThird, for each of these two peak months, extract detailed data. For each cluster, obtain the total complaint count for that month, the names of the two excluded top companies, and their combined complaint count in the original peak month. Also identify the state with the highest complaint count in that feature month and its complaint count, the most complained-about company within that state and its complaint count, and the full details of the earliest complaint received by that company in that state during that feature month (including Complaint ID, date, Issue, and Sub-issue).\n\nFinally, after completing all the above steps, format the two sets of core parameters into a single string with no spaces. The two rows must be ordered AllCompanies first, ResidualAfterTop2 second. The field order for each row is fixed as: <Cluster>|<Month>|<MonthComplaints>|<ExcludedTop2Companies>|<ExcludedTop2PeakMonthComplaints>|<TopState>|<StateComplaints>|<TopCompany>|<CompanyComplaints>|<EarliestComplaintID>|<EarliestDate>|<Issue>|<SubIssue>. To ensure a unique answer: output the two records in AllCompanies then ResidualAfterTop2 order; list ExcludedTop2Companies in descending order of their complaint count in the original peak month, breaking ties alphabetically; for any tie in sorting indicators, sort by field value in ascending alphabetical order; if the earliest complaint date is tied, select the one with the smallest Complaint ID.start_urlhttps://www.consumerfinance.gov/data-research/consumer-complaints/output_formatOutput the two rows in AllCompanies first, ResidualAfterTop2 second order, joined into a formatted string with no spaces. The field order for each row is fixed as: <Cluster>|<Month>|<MonthComplaints>|<ExcludedTop2Companies>|<ExcludedTop2PeakMonthComplaints>|<TopState>|<StateComplaints>|<TopCompany>|<CompanyComplaints>|<EarliestComplaintID>|<EarliestDate>|<Issue>|<SubIssue>.oracle_answerAllCompanies|2021-05|336|Coinbase, Inc.+BAM Management US Holdings Inc.|226|CA|58|Coinbase, Inc.|21|4341451|2021-05-01|Money was not available when promised|NONE,ResidualAfterTop2|2020-06|180|Coinbase, Inc.+BAM Management US Holdings Inc.|226|OH|28|PNC Bank N.A.|22|3687376|2020-06-07|Other transaction problem|NONEmetadata{\:[\,\,\],\:\,\:true,\:true,\:[\,\,\],\:[\,\,\],\:{\:\},\:\}rubric{\:[\,\,\],\:[\,\,\],\:{\:\,\:\,\:[\,\,\,\,\,\,\,\,\,\,\,\,\],\:[\,\],\:{\:\,\:\},\:\,\:\,\:\,\:\,\:\}}all_involved_urlsnull |
| task_idconsumerfinance_007domainCFPB_COMPLAINT_DATABASEautonomy_typeordered tableoracle_output_cardinalityinstructionI am drafting a layered analysis report on the core factors behind a batch of cryptocurrency complaints. I need your help to verify the true causes of this complaint surge step by step using the CFPB complaint database.\n\nFirst, define the base scope of the analysis. Using the current CFPB product and issue taxonomy documentation, confirm the product bucket that \ falls under. Once the bucket is determined, strictly restrict the time range to between 2020-01-01 and 2022-08-26, and only for the \ subcategory under that product bucket, extract the original peak month (BaselinePeak) and its total complaint count for this initial complaint cohort.\n\nSecond, determine the shift in peaks by progressively removing core influencing factors. Round 1: In the original peak month identified above, identify the most frequent \ value, then completely remove all complaints with that value from the entire initial cohort to obtain the recalculated second peak month (AfterDominantPublicResponse). Round 2: In the second peak month, identify the most frequent \ type, remove it from the residual data of the previous round to lock in the third emerging peak month (AfterDominantPublicResponseAndIssue). Round 3: In the third peak month, identify the single most complained-about company (Company), remove that company, and perform the final round of statistics to obtain the final residual peak month (AfterDominantPublicResponseIssueAndCompany).\n\nThird, for each of the identified characteristic months, extract detailed data. For these four complaint clusters, besides obtaining the total complaint count for that month, also determine the newly added removal dimension and its specific excluded value, and the count of that excluded value in the month that served as its reference source. Additionally, identify the state with the highest complaint count in that month and its complaint count, the most complained-about company within that state and its complaint volume, and the complete information of the earliest complaint received by the highest-complaint company in that state during that characteristic month, including complaint ID, date, Issue, and Sub-issue.\n\nFinally, after completing the investigation and information collation for all the above dimensions, concatenate all extracted parameters for these four complaint clusters strictly in the order of BaselinePeak, AfterDominantPublicResponse, AfterDominantPublicResponseAndIssue, and AfterDominantPublicResponseIssueAndCompany into four strings with no spaces. Each line's field order is strictly fixed as: <Cluster>|<Month>|<MonthComplaints>|<NewExclusionType>|<NewExclusionValue>|<NewExclusionCountInSourcePeak>|<TopState>|<StateComplaints>|<TopCompany>|<CompanyComplaints>|<EarliestComplaintID>|<EarliestDate>|<Issue>|<SubIssue>. In case of ties, always select by lexicographic ascending order; for the earliest complaint with the same date, select by ascending Complaint ID numeric value. Normalize all empty values to NONE before sorting.","start_url":"https://www.consumerfinance.gov/data-research/consumer-complaints/","output_format":"Concatenate into four strings with no spaces, strictly in the order of BaselinePeak, AfterDominantPublicResponse, AfterDominantPublicResponseAndIssue, AfterDominantPublicResponseIssueAndCompany; each line's field order is strictly fixed as: <Cluster>|<Month>|<MonthComplaints>|<NewExclusionType>|<NewExclusionValue>|<NewExclusionCountInSourcePeak>|<TopState>|<StateComplaints>|<TopCompany>|<CompanyComplaints>|<EarliestComplaintID>|<EarliestDate>|<Issue>|<SubIssue>. In case of ties, always select by lexicographic ascending order; for the earliest complaint with the same date, select by ascending Complaint ID numeric value. Normalize all empty values to NONE before sorting.oracle_answerBaselinePeak|2021-05|336|NONE|NONE|0|CA|58|Coinbase, Inc.|21|4341451|2021-05-01|Money was not available when promised|NONE,AfterDominantPublicResponse|2022-07|31|Company public response|NONE|321|CA|6|BANK OF AMERICA, NATIONAL ASSOCIATION|4|5735227|2022-07-05|Other transaction problem|NONE,AfterDominantPublicResponseAndIssue|2021-10|13|Issue|Fraud or scam|26|TX|2|BANK OF AMERICA, NATIONAL ASSOCIATION|1|4821611|2021-10-18|Other transaction problem|NONE,AfterDominantPublicResponseIssueAndCompany|2021-06|8|Company|Foris DAX, Inc.|5|NC|3|TRUIST FINANCIAL CORPORATION|2|4462160|2021-06-15|Money was not available when promised|NONEmetadata{\:[\,\,\],\:\,\:true,\:true,\:[\,\,\],\:[\,\,\],\:{\:\},\:\}rubric{\:[\,\,\],\:[\,\,\],\:{\:\,\:\,\:[\,\,\,\,\,\,\,\,\,\,\,\,\,\],\:[\,\,\,\],\:{\:\,\:\},\:\,\:\,\:\,\:\,\:\,\:\}}all_involved_urlsnull |
| task_ideuropemc_001domainEUROPEPMC_PMCautonomy_typeordered tableoracle_output_cardinalityinstructionI want to consolidate the admission variables actually used by several prediction schemes in this 2020 COVID-19 mortality risk prediction paper into a single line, to make it easy to see which variables overlap and which appear only in the comparison schemes.\nPlease follow the steps below. First, locate the target paper — a 2020 open-access review whose title contains all of \, \, and \, and whose main text discusses hospitalized patients in Wuhan.\nThen, retain only two types of variables:\n- Type 1: admission variables that appear in both the Random Forest and XGBoost variable importance rankings.\n- Type 2: additional admission variables that do not appear in both rankings but are present in the CURB-65 or NEWS schemes used by the authors for comparison.\nNext, for each retained variable, record it in the following format:\nvariable~RF_rank_or_0~XGB_rank_or_0~F~C~N~X\nWhere:\n- RF_rank_or_0: the rank of this variable in the Random Forest ranking; write 0 if it is not included.\n- XGB_rank_or_0: the rank of this variable in the XGBoost ranking used in the paper; write 0 if it is not included.\n- F: whether this variable entered the authors' final logistic regression model.\n- C: whether this variable belongs to CURB-65.\n- N: whether this variable belongs to NEWS.\n- X: whether this variable belongs to the XGBoost model used as a comparison in the paper.\nFinally, sort all variables alphabetically by variable name, join them with | into a single line; if the paper cannot be found, output NONE.","start_url":"https://europepmc.org/advancesearch","output_format":"Next, for each retained variable, record it in the following format:\nvariable~RF_rank_or_0~XGB_rank_or_0~F~C~N~X\nWhere:\n- RF_rank_or_0: the rank of this variable in the Random Forest ranking; write 0 if it is not included.\n- XGB_rank_or_0: the rank of this variable in the XGBoost ranking used in the paper; write 0 if it is not included.\n- F: whether this variable entered the authors' final logistic regression model.\n- C: whether this variable belongs to CURB-65.\n- N: whether this variable belongs to NEWS.\n- X: whether this variable belongs to the XGBoost model used as a comparison in the paper.\nFinally, sort all variables alphabetically by variable name, join them with | into a single line; if the paper cannot be found, output NONE.oracle_answerAge~4~2~1~1~0~0|Alanine aminotransferase~17~11~0~0~0~0|Aspartate transaminase~6~4~0~0~0~0|Blood urea nitrogen~5~7~0~1~0~0|Confusion~0~0~0~1~1~0|CRP~2~6~1~0~0~1|Creatinine~9~13~0~0~0~0|Diastolic pressure~16~17~0~1~0~0|Heart rate~15~14~0~0~1~0|LDH~1~1~1~0~0~1|Lymphocyte count~3~5~0~0~0~1|Monocytes~12~16~0~0~0~0|Neutrophil~7~3~0~0~0~0|Normal platelet~10~9~0~0~0~0|Oxygen inhalation~0~0~0~0~1~0|Respiratory rate~11~10~0~1~1~0|SPO2~0~0~0~0~1~0|Systolic pressure~13~12~0~1~1~0|Temperature~18~18~0~0~1~0|Total bilirubin~14~15~0~0~0~0|White blood cell~8~8~0~0~0~0metadata{\:[\prognosis model\\\mortality risk\\\COVID-19\\\,\,\],\:\,\:true,\:true,\:[\,\,\],\:[\,\,\],\:{\:\},\:\}rubric{\:[\prognosis model\\\mortality risk\\\COVID-19\\\,\,\],\:[\,\,\],\:{\:\,\:\,\:[\,\,\,\,\,\,\],\:\,\:\,\:\,\:\,\:[\,\]}}all_involved_urlsnull |
| task_ideuropemc_002domainEUROPEPMC_PMCautonomy_typeordered tableoracle_output_cardinalityinstructionTrack studies that were Phase 1 in the 2021 Alzheimer's disease drug development pipeline review by their ClinicalTrials.gov identifier, and check whether they continue to appear in the 2022 and 2023 pipeline reviews.\nFollow these steps. First, locate the three articles: the 2021 review titled \"Alzheimer's disease drug development pipeline\,:,:,:,:State-Gated Retrieval\Use only the Phase 1 studies from the 2021 Alzheimer's disease drug development pipeline article as the baseline.\",\"Retention in 2022 and 2023 must be determined by exact match on the ClinicalTrials.gov NCT identifier, not by drug name alone.\",\"Keep only studies that are found under the same NCT in both 2022 and 2023, and output them sorted by NCT in ascending order.\"],\"dependency_type\":\"Data + Control\",\"intra_chain\":true,\"inter_chain\":true,\"data_dependency\":[\"the 2021, 2022, and 2023 pipeline reviews yield the 2021 Phase 1 baseline cohort and the later-year lookup tables\",\"article tables yield exact ClinicalTrials.gov NCT IDs, yearly phase labels, and yearly drug names for the same study\",\"final rows are produced only after 2021 baseline studies are re-identified in both later reviews by exact NCT matching\"],\"control_dependency\":[\"continuation must be matched by exact NCT rather than by drug name, because drug names drift across years\",\"2022 and 2023 lookups must use the fixed 2021 Phase 1 NCT set as the starting cohort rather than later-year table membership alone\",\"yearly phase/name fields are only valid after the same NCT is confirmed across all three reviews\"],\"freeze\":{\"historical_window\":\"the 2021, 2022, and 2023 open-access Alzheimer's disease drug development pipeline reviews; the baseline is the 2021 Phase 1 cohort and continuation is matched by exact ClinicalTrials.gov NCT\answer_type\multi-row ordered table\,:inclusion_conditions\Use only the Phase 1 studies from the 2021 Alzheimer's disease drug development pipeline article as the baseline.\",\"Retention in 2022 and 2023 must be determined by exact match on the ClinicalTrials.gov NCT identifier, not by drug name alone.\",\"Keep only studies that are found under the same NCT in both 2022 and 2023, and output them sorted by NCT in ascending order.\"],\"exclusion_conditions\":[\"Exclude results that are continued by drug name alone without an exact match on the same ClinicalTrials.gov identifier.\",\"Exclude studies found only in 2022 or only in 2023, not continuously retained in both years.\",\"Exclude studies not in the 2021 baseline Phase 1 set.\"],\"normalization\":{\"field_separator\":\"~\",\"record_separator\":\"|\",\"schema\":[\"NCT\",\"agent_2021\",\"phase_2022\",\"agent_2022\",\"phase_2023\",\"agent_2023\"],\"NCT\":\"use the exact ClinicalTrials.gov identifier as the cross-year identity key\",\"phase_fields\":\"normalize phases to P1, P2, or P3; if the same NCT appears multiple times in a year, keep the highest phase for that year\",\"agent_fields\":\"keep the agent text as shown in each year-specific table\",\"sorting_or_selection\":\"NCT ascending\"}}","all_involved_urls":"null"} |
| {"task_id":"europemc_003","domain":"EUROPEPMC_PMC","autonomy_type":"ordered table","oracle_output_cardinality":8,"instruction":"I want to see which candidate molecules predicted in the 2023 \"Antibodies to Watch\" article as expected to submit their first marketing applications in 2022–2023 have actually been approved or entered first regulatory review in the 2024 installment of the same series.\nFollow the steps below. First, find the open-access full texts of both the 2023 and 2024 \"Antibodies to Watch\" articles in Europe PMC.\nThen, in the 2023 article, identify all candidate molecules listed as **\"expected to submit first marketing applications in 2022–2023\"**.\nNext, check each of these molecules against the 2024 article and retain only those that have been approved or have entered first regulatory review.\nFinally, format each qualifying molecule as follows:\nINN~2023 category~2023 prediction label~2024 status~2024 indication~2024 region\nwhere:\n\t•\t2023 category uses NC/CA for non-cancer/cancer;\n\t•\t2024 status uses A/R for first approval/first regulatory review.\nSort by INN alphabetically and output as a single line joined by |; if neither article is found, output NONE.","start_url":"https://europepmc.org/advancesearch","output_format":"Finally, format each qualifying molecule as follows:\nINN~2023 category~2023 prediction label~2024 status~2024 indication~2024 region\nwhere:\n\t•\t2023 category uses NC/CA for non-cancer/cancer;\n\t•\t2024 status uses A/R for first approval/first regulatory review.\nSort by INN alphabetically and output as a single line joined by |; if neither article is found, output NONE.","oracle_answer":"Cosibelimab~CA~MAA,Q2 2023~R~Squamous cell carcinoma~US|Elranatamab~CA~2023~A~Multiple myeloma~US,EU,Switzerland,Brazil|Garadacimab~NC~BLA,2023~R~Hereditary angioedema~EU|Odronextamab~CA~BLA,H2 2022~R~Diffuse large B-cell lymphoma~US,EU|Pozelimab~NC~BLA,H2 2022~A~CHAPLE disease~US|Suciraslimab~NC~NDA,2023~R~Rheumatoid arthritis~China|Talquetamab~CA~2023–25~A~Multiple myeloma~US,EU,UK,Switzerland|Zolbetuximab~CA~BLA,2023~R~HER2-negative gastric or gastroesophageal junction adenocarcinoma~US,EU,Japan,China","metadata":"{\"State-Gated Retrieval\":[\"First, identify all candidate molecules in the 2023 Antibodies to Watch article that are predicted to submit first marketing applications in 2022-2023.\",\"Then, check each of these molecules against the 2024 article and retain only those that have been first approved or have entered first regulatory review.\",\"The 2024 status must distinguish A=first approval from R=first regulatory review, and include the 2024 indication and region.\"],\"dependency_type\":\"Data + Control\",\"intra_chain\":true,\"inter_chain\":true,\"data_dependency\":[\"the 2023 Antibodies to Watch article yields the forecast baseline of candidates expected to file in 2022-2023\",\"the 2024 article tables yield first-approval outcomes, first-regulatory-review outcomes, indications, and regions for the same INNs\",\"final rows are produced only after Table 1 approvals and Table 2 first-review candidates are merged back onto the 2023 forecast cohort\"],\"control_dependency\":[\"both 2024 Table 1 and Table 2 must be read so that first-review-but-not-yet-approved candidates are retained\",\"2024 status coding must distinguish approval from first regulatory review instead of collapsing both into a single positive match\",\"region and indication fields are valid only after the 2023 forecast cohort is reconciled against both 2024 tables\"],\"freeze\":{\"historical_window\":\"the 2023 and 2024 open-access Antibodies to Watch articles; baseline candidates come from the 2023 forecast set and 2024 outcomes come from current Table 1 plus Table 2\"},\"answer_type\":\"multi-row ordered table\"}","rubric":"{\"inclusion_conditions\":[\"First, identify all candidate molecules in the 2023 Antibodies to Watch article that are predicted to submit first marketing applications in 2022-2023.\",\"Then, check each of these molecules against the 2024 article and retain only those that have been first approved or have entered first regulatory review.\",\"The 2024 status must distinguish A=first approval from R=first regulatory review, and include the 2024 indication and region.\"],\"exclusion_conditions\":[\"Exclude results that only check approvals in 2024 Table 1 without also consulting Table 2 for first regulatory review candidates.\",\"Exclude molecules not in the 2023 forecast baseline.\",\"Exclude candidates that were neither first approved nor entered first regulatory review in 2024.\"],\"normalization\":{\"field_separator\":\"~\",\"record_separator\":\"|\",\"schema\":[\"INN\",\"category_2023\",\"prediction_label_2023\",\"status_2024\",\"indication_2024\",\"region_2024\"],\"category_2023\":\"emit `NC` or `CA` exactly\",\"prediction_label_2023\":\"keep the label content inside the parentheses and remove spaces after commas, for example `BLA,H2 2022`\",\"status_2024\":\"emit `A` for first approval and `R` for first regulatory review\",\"region_2024\":\"preserve the raw region string but remove spaces after commas, for example `US,EU`\",\"sorting_or_selection\":\"INN alphabetical ascending\"}}","all_involved_urls":"null"} |
| {"task_id":"europemc_004","domain":"EUROPEPMC_PMC","autonomy_type":"ordered table","oracle_output_cardinality":9,"instruction":"I am evaluating the accuracy of annual antibody drug watchlist predictions and want to trace whether the 2022 cohort of candidates from the Antibodies to Watch series simply experienced delays in development progress. I need you to use the Europe PMC database to systematically map the year-over-year status changes for these candidates.\n\nFirst, establish the literature baseline and initial candidate set. Precisely locate the three open-access full-text Antibodies to Watch articles published in 2022, 2023, and 2024 in Europe PMC. Once all three articles are found, identify a set of antibodies (INN) with a specific delay pattern: those that were explicitly listed in the 2022 article under a category such as \"expected to file soon\" (identified by INN, drug code, project name, or other recognizable name) but, due to slower progress, still remained as watchlist candidates in the 2023 article.\n\nSecond, determine the latest status of these consecutively delayed candidates. For the identified antibody group, ascertain their final status in the 2024 article: whether they finally received first approval, are in first regulatory review, remain on the 2024 watchlist, or have completely disappeared from the 2024 main article list.\n\nFinally, after completing the cross-year literature review and status confirmation, aggregate the extracted parameters into a specific formatted single-line record. If all three target articles cannot be found, output NONE. If all three are successfully located, concatenate the internal parameters for each qualifying antibody in the fixed order \"INN~2022 category~2022 phase~2023 prediction label~2024 status~2024 phase or NA~2024 indication or NA~2024 region or NA\". Then join all identified candidates in strict alphabetical order by INN using the \"|\" symbol into a single string with no spaces.","start_url":"https://europepmc.org/advancesearch","output_format":"If all three target articles cannot be found, output NONE. If all three are successfully located, concatenate the internal parameters for each qualifying antibody in the fixed order \"INN~2022 category~2022 phase~2023 prediction label~2024 status~2024 phase or NA~2024 indication or NA~2024 region or NA\". Then, join all identified candidates in strict alphabetical order by INN using the \"|\" symbol into a single string with no spaces.","oracle_answer":"Apamistamab-I-131~CA~P3~BLA,H1 2023~F~P3~Ablation of bone marrow prior to transplantation in AML patients~NA|Bentracimab~NC~P3~BLA,Q4 2022~F~P3~Reversal of the antiplatelet effects of ticagrelor~NA|Cosibelimab~CA~P3~MAA,Q2 2023~R~NA~Squamous cell carcinoma~US|Erfonrilimab~CA~P3~NDA~F~P3~Pancreatic ductal adenocarcinoma~NA|Odronextamab~CA~P2~BLA,H2 2022~R~NA~Diffuse large B-cell lymphoma~US,EU|Talquetamab~CA~P2~2023-25~A~NA~Multiple myeloma~US,EU,UK,Switzerland|Tiragolumab~CA~P3~2023~F~P3~NSCLC~NA|Zanidatamab~CA~P2~2023~D~NA~NA~NA|Zolbetuximab~CA~P3~BLA,2023~R~NA~HER2-negative gastric or gastroesophageal junction adenocarcinoma~US,EU,Japan,China","metadata":"{\"State-Gated Retrieval\":[\"Start with the 2022 Antibodies to Watch candidates listed as \\\"expected to file soon\\\", but retain only those molecules that still remain as watchlist candidates in the 2023 article of the same series.\",\"For these consecutively delayed candidates, determine their final status in the 2024 article: first approved, in first regulatory review, or completely absent from Tables 1-4.\",\"Name matching must handle the Apamistamab naming drift; complete absence in 2024 must be explicitly encoded as D.\"],\"dependency_type\":\"Data + Control\",\"intra_chain\":true,\"inter_chain\":true,\"data_dependency\":[\"the 2022, 2023, and 2024 Antibodies to Watch articles yield the 2022 soon-to-file baseline, 2023 carry-over membership, and 2024 status tables\",\"article tables and full-text naming context yield name normalizations, 2024 table presence/absence, and final status labels for each carried-over INN\",\"final rows are produced only after 2022 candidates are filtered by 2023 carry-over status and then resolved against 2024 Tables 1-4 plus explicit absence\"],\"control_dependency\":[\"2022 candidates must first be filtered by 2023 carry-over status before they are matched against the 2024 tables\",\"Apamistamab naming drift forces explicit normalization before 2024 lookup\",\"complete 2024 absence must be emitted as D rather than treated as a lookup failure\"],\"freeze\":{\"historical_window\":\"the 2022, 2023, and 2024 open-access Antibodies to Watch articles; the 2022 baseline is filtered by 2023 carry-over status and then checked against 2024 Tables 1-4, with explicit absence retained as a status\"},\"answer_type\":\"multi-row ordered table\"}","rubric":"{\"inclusion_conditions\":[\"Start with the 2022 Antibodies to Watch candidates listed as \\\"expected to file soon\\\", but retain only those molecules that still remain as watchlist candidates in the 2023 article of the same series.\",\"For these consecutively delayed candidates, determine their final status in the 2024 article: first approved, in first regulatory review, or completely absent from Tables 1-4.\",\"Name matching must handle the Apamistamab naming drift; complete absence in 2024 must be explicitly encoded as D.\"],\"exclusion_conditions\":[\"Exclude results that directly match the 2022 baseline against 2024 without applying the 2023 carry-over filter.\",\"Exclude matching results that do not handle the Apamistamab naming drift.\",\"Exclude cases where complete 2024 absence is treated as a lookup failure rather than a status that must be output.\"],\"normalization\":{\"field_separator\":\"~\",\"record_separator\":\"|\",\"schema\":[\"INN\",\"category_2022\",\"phase_2022\",\"prediction_label_2023\",\"status_2024\",\"phase_or_NA_2024\",\"indication_or_NA_2024\",\"region_or_NA_2024\"],\"category_2022\":\"emit `NC` or `CA` exactly\",\"phase_2022\":\"normalize stage labels such as `Phase 3` to `P3` and `Pivotal Phase 2` to `P2`\",\"prediction_label_2023\":\"keep the label content inside the parentheses and remove spaces after commas, for example `BLA,H1 2023`\",\"status_2024\":\"emit one of `A`, `R`, `F`, or `D` exactly\",\"conditional_fields_2024\":\"if status is A or R, phase_or_NA_2024 must be `NA` and region_or_NA_2024 must carry the raw region string; if status is F, region_or_NA_2024 is `NA` and phase_or_NA_2024 carries the normalized phase; if status is D, all three trailing 2024 fields are `NA`\",\"naming_exception\":\"preserve the fixed source exception around the Apamistamab-I-131 naming variant instead of over-normalizing the INN\",\"sorting_or_selection\":\"INN alphabetical ascending\"}}","all_involved_urls":"null"} |
| {"task_id":"europemc_005","domain":"EUROPEPMC_PMC","autonomy_type":"ordered table","oracle_output_cardinality":20,"instruction":"I am compiling a cross-platform correspondence checklist for core predictors of COVID-19 hospitalization prognosis. Using the Europe PMC database, I need to work through the process of identifying the variables each specific model actually used, together with their matching attributes.\n\nFirst, establish the literature baseline. In Europe PMC, precisely locate the 2024 WHO global individual patient data validation paper and identify the four prognostic models that were successfully externally validated in low- and middle-income country data. Then trace back to the three original model papers corresponding to these four models.\n\nAfter locating these four target papers, the core task is to extract and verify the specific characteristics of the predictors. Determine which distinct admission variables each of the four models actually used in the evaluation. For each identified variable, clarify its acquisition pathway in the WHO platform baseline table: either a direct match under the same or a substantively equivalent name (marked as E), or a fixed mapped conversion (marked as M). Also trace the model-source attributes of these variables: whether they belong to the model with the strongest external validation performance (TopAUC), the model with calibration closest to the ideal (BestCal), or an original model that explicitly used dynamic data (Dyn).\n\nTo ensure consistent judgment criteria, model attribution, performance attribution, and variable conflicts must all be based on the verifiable model descriptions, performance summaries, and field definitions in the target papers. When the same concept has multiple naming or mapping methods, prioritize conservative mappings that maintain cross-model interpretability consistency, and flag mapping differences that may affect interpretability consistency.\n\nAfter completing the above literature tracing and variable characteristic extraction, finally extract these parameters to generate a formatted single-line summary record. For status flags, uniformly use 1 for yes and 0 for no for all boolean types. If the search fails to locate the one validation paper and the three original model papers, output NONE. If all are successfully located, after arranging all distinct variables strictly in alphabetical order by variable name, concatenate the parameters within each variable in the fixed order \"variable~category~B~W~Zh~Zo~WHOfield~WHOx~TopAUC~BestCal~Dyn~Conflict\", then join all variables with \"|\" into a single string with no spaces.","start_url":"https://europepmc.org/advancesearch","output_format":"If the search fails to locate the one validation paper and the three original model papers, output NONE. If all are successfully located, after arranging all distinct variables strictly in alphabetical order by variable name, concatenate the parameters within each variable in the fixed order \"variable~category~B~W~Zh~Zo~WHOfield~WHOx~TopAUC~BestCal~Dyn~Conflict\", then join all variables with \"|\" into a single string with no spaces. For all boolean types, uniformly use 1 for yes and 0 for no.","oracle_answer":"Age~DEM~1~1~1~1~Age~E~1~1~1~0|Cancer~COM~0~0~1~0~Cancer~E~1~0~0~0|Chronic kidney disease~COM~0~0~0~1~CKD~M~0~0~0~1|Chronic lung disease~COM~0~0~1~0~Pulmonary disease~M~1~0~0~0|Chronic renal disease~COM~0~0~1~0~CKD~M~1~0~0~1|Coronary heart disease~COM~0~1~0~0~CHD~M~0~0~0~1|Cough~SYM~0~0~1~0~Cough~E~1~0~0~0|Diabetes mellitus~COM~0~0~1~0~Diabetes~M~1~0~0~0|Diarrhea~SYM~0~0~1~0~Diarrhea~E~1~0~0~0|Dyspnea~SYM~0~0~1~0~Dyspnea~E~1~0~0~0|Fever~VIT~0~0~0~1~Body temperature~M~0~0~0~0|Heart disease~COM~0~0~1~0~CHD~M~1~0~0~1|Hypertension~COM~0~1~1~0~Hypertension~E~1~0~0~0|Immunocompromised status~COM~0~0~1~0~Immunosuppression~M~1~0~0~0|O2 saturation~VIT~1~0~0~0~Oxygen saturation~M~0~1~1~0|Respiratory rate~VIT~1~0~0~1~Respiratory rate~E~0~1~1~0|Sex~DEM~0~0~1~0~Men~M~1~0~0~0|Smoking~DEM~1~0~0~0~Smoking~E~0~1~1~1|Smoking status~DEM~0~0~0~1~Smoking~M~0~0~0~1|Systolic blood pressure~VIT~0~0~0~1~Systolic blood pressure~E~0~0~0~0","metadata":"{\"State-Gated Retrieval\":[\"First, in the 2024 WHO global individual patient data validation paper, identify the four prognostic models that were successfully externally validated in LMIC data, and trace back to the three corresponding original model papers.\",\"For each predictor, retain the original wording from WHO Table 1, annotate its matching method (E/M) in the WHO baseline table, and indicate whether it comes from the TopAUC, BestCal, or Dyn model.\",\"When encountering a predictor that was conceptually merged on the WHO side, trace back to the original Wang/Zhou papers to split it apart and then recalculate the relevant fields.\"],\"dependency_type\":\"Data + Control\",\"intra_chain\":true,\"inter_chain\":true,\"data_dependency\":[\"the 2024 WHO global validation paper yields the four externally validated models and the WHO-side predictor wording/mapping baseline\",\"the three original model papers yield model-specific predictor wordings, dynamic-data flags, and conflicts hidden by WHO-side conceptual merging\",\"final rows are produced only after WHO predictors are kept separate, traced back to the original papers, and then re-annotated for mapping and model provenance\"],\"control_dependency\":[\"conceptually merged WHO predictors must be split back out when the original Wang/Zhou papers expose distinct variables\",\"Table 2 mapping decisions are not valid until Table 1 wording and original-paper predictor sets are reconciled\",\"TopAUC / BestCal / Dyn flags must be determined after the predictor split is applied correctly\"],\"freeze\":{\"historical_window\":\"the 2024 WHO global IPD validation paper and the three linked original COVID prognosis-model papers, using their current open-access full texts and tables\"},\"answer_type\":\"multi-row ordered table\"}","rubric":"{\"inclusion_conditions\":[\"First, in the 2024 WHO global individual patient data validation paper, identify the four prognostic models that were successfully externally validated in LMIC data, and trace back to the three corresponding original model papers.\",\"For each predictor, retain the original wording from WHO Table 1, annotate its matching method (E/M) in the WHO baseline table, and indicate whether it comes from the TopAUC, BestCal, or Dyn model.\",\"When encountering a predictor that was conceptually merged on the WHO side, trace back to the original Wang/Zhou papers to split it apart and then recalculate the relevant fields.\"],\"exclusion_conditions\":[\"Exclude results that directly merged conceptually similar predictors after WHO Table 1 without tracing back to the original papers for splitting.\",\"Exclude results that relied solely on loose mapping from WHO Table 2 without incorporating the original paper wording.\",\"Exclude results that fixed TopAUC/BestCal/Dyn annotations before applying the predictor split correction.\"],\"normalization\":{\"field_separator\":\"~\",\"record_separator\":\"|\",\"schema\":[\"variable\",\"category\",\"B\",\"W\",\"Zh\",\"Zo\",\"WHOfield\",\"WHOx\",\"TopAUC\",\"BestCal\",\"Dyn\",\"Conflict\"],\"variable_set\":\"the candidate variable set comes only from the WHO validation Table 1 and remains at the original 20 predictors\",\"conflict_groups\":\"do not pre-merge the three source conflict groups `Smoking`/`Smoking status`, `Chronic renal disease`/`Chronic kidney disease`, and `Coronary heart disease`/`Heart disease` before applying the conflict flag\",\"boolean_fields\":\"B, W, Zh, Zo, TopAUC, BestCal, Dyn, and Conflict are emitted as 1 or 0 only\",\"WHOx\":\"emit `E` for exact WHO matches or direct matches under a substantively equivalent name, `M` for the fixed mapped match, and `NA` when no WHO field matches\",\"sorting_or_selection\":\"variable name alphabetical ascending\"}}","all_involved_urls":"null"} |
| {"task_id":"genome_001","domain":"KEGG_BRITE","autonomy_type":"ordered table","oracle_output_cardinality":3,"instruction":"I am compiling a reference list of medications for ulcerative colitis for in-hospital use. From the immunosuppressants cataloged in KEGG, I want to select those drug entries that have been approved in Japan by the end of June 2025 and that carry explicit CYP-related metabolism information or CYP inhibition indications. Please proceed in the following order: first, within the scope of KEGG immunosuppressants for ulcerative colitis, screen for drug entries approved in Japan by the end of June 2025; then check whether each entry has explicit CYP-related metabolism information or CYP inhibition indications, retaining only qualifying entries. Next, record each entry's Japanese therapeutic classification code, ATC code, corresponding target branch, and Japan first approval date. Finally, output a single-line string sorted by Japan first approval date from earliest to latest, in the format <DRUG_ID>|<JP_class>|<ATC>|<Target_branch>|<JP_date>; separate entries with commas; if no qualifying entries exist, output NONE.start_urlhttps://www.genome.jp/kegg/brite.htmloutput_formatOutput a single-line string sorted by Japan first approval date from earliest to latest, in the format <DRUG_ID>|<JP_class>|<ATC>|<Target_branch>|<JP_date>; separate entries with commas; if no qualifying entries exist, output NONE.oracle_answerD09783|3999|L04AF01|JAK1/JAK2/JAK3/TYK2|2013/3/25,D11048|3999|L04AF03|JAK1|2020/1/23,D10931|2399|L04AE05|S1PR1/S1PR4/S1PR5|2025/6/24metadata{\:[\,\,\],\:\,\:true,\:true,\:[\,\,\],\:[\,\,\],\:{\:\},\:\}rubric{\:[\,\,\],\:[\,\,\],\:{\:\,\:\,\:[\,\,\,\,\],\:\,\:\,\:\,\:{\:\,\:\}}}all_involved_urlsnull |
| task_idgenome_002domainKEGG_GENOMEautonomy_typeordered tableoracle_output_cardinalityinstructionI am preparing a lecture page on classic animal endosymbiosis cases and want to find, among KEGG's host-symbiont examples, the case that meets the following conditions: it was already recorded by the end of 2005, its bacterial symbiont genome contains named extra replicons, and the names of those replicons directly correspond to amino acid synthesis modules on the host-symbiont joint pathway map. Process in the following order: first, among KEGG's host-symbiont examples, select those combinations whose bacterial symbiont KEGG GENOME entry has a Created year on or before 2005; then retain cases where the bacterial symbiont genome carries named extra replicons; next, verify that the replicon names directly correspond to amino acid synthesis modules on the host-symbiont joint pathway map; finally, within the joint pathway map for that host-symbiont combination, list all qualifying pathway-module pairs for that case and output a single-line string: <group>|<symbiont>|<pathway>|<module>|<replicon>|<created_year> (sorted by pathway ID in ascending order, separated by commas; output NONE if no qualifying items exist).start_urlhttps://www.genome.jp/kegg/genome/output_formatOutput a single-line string: <group>|<symbiont>|<pathway>|<module>|<replicon>|<created_year> (sorted by pathway ID in ascending order, separated by commas; output NONE if no qualifying items exist).oracle_answerapi+buc|buc|00290|M00432|pLeu|2000,api+buc|buc|00400|M00023|pTrp|2000metadata{\:[\,\,\],\:\,\:true,\:true,\:[\,\,\],\:[\,\,\],\:{\:\},\:\}rubric{\:[\,\,\],\:[\,\,\],\:{\:\,\:\,\:[\,\,\,\,\,\],\:\,\:\,\:\,\:\,\:\,\:\,\:\}}all_involved_urlsnull |
| task_idgenome_003domainKEGG_BRITEautonomy_typeordered tableoracle_output_cardinalityinstructionI am compiling a table of the approval order of systemic drugs for plaque psoriasis in Japan. Process as follows: first, within the scope of systemic drugs for plaque psoriasis approved in Japan by the end of 2024, identify the active ingredients that act on the IL-17/IL-23 signaling axis; then exclude all topical drugs and systemic drugs acting on other axes. Next, sort these active ingredients by their first approval date in Japan from earliest to latest; if approved on the same day, sort by DRUG_ID in ascending order. Finally, output a multi-row table with each row formatted as <DRUG_ID>|<Name>|<ATC>|<Target_branch>|<PMDA_date>.start_urlhttps://www.kegg.jp/brite/br08301output_formatOutput a multi-row table with each row formatted as <DRUG_ID>|<Name>|<ATC>|<Target_branch>|<PMDA_date>.oracle_answerD09967|Secukinumab|L04AC10|IL17A|2014/12/26\nD10061|Brodalumab|L04AC12|IL17RA|2016/7/4\nD10071|Ixekizumab|L04AC13|IL17A|2016/7/4\nD10438|Guselkumab|L04AC16|IL23A|2018/3/23\nD11052|Risankizumab|L04AC18|IL23A|2019/3/26\nD10400|Tildrakizumab|L04AC17|IL23A|2020/6/29\nD11550|Bimekizumab|L04AC21|IL17A/IL17F|2022/1/20\nD11817|Deucravacitinib|L04AF07|TYK2|2022/9/26metadata{\:[\,\,\],\:\,\:true,\:true,\:[\,\,\],\:[\,\,\],\:{\:\},\:\}rubric{\:[\,\,\],\:[\,\,\],\:{\:\,\:\,\:\,\:[\,\,\,\,\],\:\,\:\,\:{\:\,\:\}}}all_involved_urlsnull |
| task_idgenome_004domainKEGG_GENOMEautonomy_typeordered tableoracle_output_cardinalityinstructionI am preparing a handout for a primate evolution course and want to summarize, in a two-row comparison table, the anomalous pattern behind why this lineage must obtain vitamin C from the diet. Among primate genome entries established by the end of 2022, identify the anomalous lower-level clade in which the upstream glucuronate module M00014 is still retained but the animal-type vitamin C biosynthesis module M00129 has been lost. Then verify the corresponding sister branch and confirm that it still retains M00129. Finally, output a two-row comparison table with columns rank|clade_role|parent_clade|clade|anomaly_pathway|contrast_pathway|retained_upstream_module|vitamin_c_synthesis_module|vitamin_c_module_status|example_org|cutoff_year. The first row should be EXCEPTION_CLADE and the second row RETAINED_SISTER_CLADE. In all rows, vitamin_c_synthesis_module is fixed as M00129; vitamin_c_module_status should be LOST for the first row and RETAINED for the second. Use KEGG organism codes for the example species.","start_url":"https://www.genome.jp/kegg/kegg2.html","output_format":"Output a two-row comparison table with columns rank|clade_role|parent_clade|clade|anomaly_pathway|contrast_pathway|retained_upstream_module|vitamin_c_synthesis_module|vitamin_c_module_status|example_org|cutoff_year.","oracle_answer":"1|EXCEPTION_CLADE|Primates|Haplorrhini|map00053|map00040|M00014|M00129|LOST|csyr|2022\n2|RETAINED_SISTER_CLADE|Primates|Strepsirrhini|map00053|map00040|M00014|M00129|RETAINED|oga|2022metadata{\:[\,\,\],\:\,\:true,\:true,\:[\,\,\],\:[\,\,\],\:{\:\},\:\}rubric{\:[\,\,\],\:[\,\,\],\:{\:\,\:\,\:\,\:[\,\,\,\,\,\,\,\,\,\,\],\:\,\:[\,\],\:\,\:\,\:[\,\],\:\,\:\,\:\}}all_involved_urlsnull |
| task_idnvd_001domainNVDautonomy_typeordered tableinstructionWe are cleaning up a batch of old SBOM matching rules for the Duplicator product line. Some rules still rely on legacy vendor lineage and outdated product labeling, so we need to identify Duplicator-related CVEs whose current product line belongs to awesomemotive:duplicator but can still be traced back to snapcreek:duplicator in the deprecated vendor chain. Search NVD records for Duplicator-related CVEs published from 2018-01-01 through 2024-12-31. Retain only CVEs that satisfy all of the following conditions in Change History: the Initial Analysis is not yet lite or lite+pro; the first change to lite or lite+pro occurs in a NIST analysis event on 2021-10-18; and a snapcreek -> awesomemotive vendor remap occurs on 2026-02-02. Exclude CVEs whose Initial Analysis already contained lite/lite+pro, as well as records that lack the required vendor-remap event or do not belong to the Duplicator product line. Finally, output a single-line ordered table with columns: CVE | lite_upper_version | lite_vs_lite+pro | analysis_event@date | snapcreek_to_awesomemotive@date.","start_url":"https://nvd.nist.gov/vuln/search#/nvd/home?resultType=records","output_format":"Output a single-line ordered table with columns: CVE | lite_upper_version | lite_vs_lite+pro | analysis_event@date | snapcreek_to_awesomemotive@date. Join rows with commas and sort by NVD published date ascending, then by CVE ascending.","oracle_output_cardinality":3,"oracle_answer":"CVE:2018-7543|1.2.32|lite|Reanalysis@2021-10-18|snapcreek->awesomemotive@2026-02-02,CVE:2018-17207|1.2.42|lite|Reanalysis@2021-10-18|snapcreek->awesomemotive@2026-02-02,CVE:2020-11738|1.3.28|lite+pro|ModifiedAnalysis@2021-10-18|snapcreek->awesomemotive@2026-02-02metadata{\:[\,\,\],\:\,\:true,\:true,\:[\,\,\],\:[\,\,\],\:{\:\},\:\}rubric{\:[\,\,\],\:[\,\,\],\:{\:\,\:\,\:\,\:[\,\,\,\,\],\:\,\:\,\:\,\:\,\:{\:\,\:\}}}all_involved_urlsnull |
| task_idnvd_002instructionIdentify which CPE 2.3 product identifiers using the legacy vendor name mysql:mysql within the MySQL 5.1 branch were explicitly migrated by NIST to the corresponding oracle:mysql CPE 2.3 product identifiers in the NVD (National Vulnerability Database) records for the four InnoDB denial-of-service vulnerabilities CVE-2010-3676, CVE-2010-3677, CVE-2010-3678, and CVE-2010-3680. Migration here refers only to the vendor field in the CPE changing from the old vendor mysql to the new vendor oracle; it does not include cases where the identifier remains under the old vendor name and only version aliases are normalized. Retain only entries from the MySQL 5.1 branch that meet this criterion. Exclude entries from the MySQL 5.0 branch, as well as all entries that did not undergo a vendor migration and only had internal version-alias normalization under the old vendor. Output a multi-row ordered table with columns version_update | legacy_cpe | current_cpe | remap_time_nvd, sorted by numeric version in ascending order; within the same version, entries without an update suffix come before entries with an update suffix.","domain":"NVD","autonomy_type":"ordered table","oracle_output_cardinality":44,"start_url":"https://nvd.nist.gov/products/cpe/search","output_format":"Output a multi-row ordered table with columns version_update | legacy_cpe | current_cpe | remap_time_nvd, sorted by numeric version in ascending order; within the same version, entries without an update suffix come before entries with an update suffix.","oracle_answer":"version_update|legacy_cpe|current_cpe|remap_time_nvd\n5.1.1|cpe:2.3:a:mysql:mysql:5.1.1:::::::*|cpe:2.3:a:oracle:mysql:5.1.1:::::::*|2019-12-17 15:04:33\n5.1.2|cpe:2.3:a:mysql:mysql:5.1.2:::::::*|cpe:2.3:a:oracle:mysql:5.1.2:::::::*|2019-12-17 15:04:39\n5.1.3|cpe:2.3:a:mysql:mysql:5.1.3:::::::*|cpe:2.3:a:oracle:mysql:5.1.3:::::::*|2019-12-17 15:04:46\n5.1.4|cpe:2.3:a:mysql:mysql:5.1.4:::::::*|cpe:2.3:a:oracle:mysql:5.1.4:::::::*|2019-12-17 15:04:53\n5.1.10|cpe:2.3:a:mysql:mysql:5.1.10:::::::*|cpe:2.3:a:oracle:mysql:5.1.10:::::::*|2019-12-17 15:05:30\n5.1.11|cpe:2.3:a:mysql:mysql:5.1.11:::::::*|cpe:2.3:a:oracle:mysql:5.1.11:::::::*|2019-12-17 15:05:37\n5.1.12|cpe:2.3:a:mysql:mysql:5.1.12:::::::*|cpe:2.3:a:oracle:mysql:5.1.12:::::::*|2019-12-17 15:05:45\n5.1.13|cpe:2.3:a:mysql:mysql:5.1.13:::::::*|cpe:2.3:a:oracle:mysql:5.1.13:::::::*|2019-12-17 15:05:52\n5.1.14|cpe:2.3:a:mysql:mysql:5.1.14:::::::*|cpe:2.3:a:oracle:mysql:5.1.14:::::::*|2019-12-17 15:05:59\n5.1.15|cpe:2.3:a:mysql:mysql:5.1.15:::::::*|cpe:2.3:a:oracle:mysql:5.1.15:::::::*|2019-12-17 15:06:07\n5.1.16|cpe:2.3:a:mysql:mysql:5.1.16:::::::*|cpe:2.3:a:oracle:mysql:5.1.16:::::::*|2019-12-17 15:06:14\n5.1.17|cpe:2.3:a:mysql:mysql:5.1.17:::::::*|cpe:2.3:a:oracle:mysql:5.1.17:::::::*|2019-12-17 15:06:21\n5.1.18|cpe:2.3:a:mysql:mysql:5.1.18:::::::*|cpe:2.3:a:oracle:mysql:5.1.18:::::::*|2019-12-17 15:06:27\n5.1.19|cpe:2.3:a:mysql:mysql:5.1.19:::::::*|cpe:2.3:a:oracle:mysql:5.1.19:::::::*|2019-12-17 15:06:34\n5.1.20|cpe:2.3:a:mysql:mysql:5.1.20:::::::*|cpe:2.3:a:oracle:mysql:5.1.20:::::::*|2019-12-17 15:06:42\n5.1.21|cpe:2.3:a:mysql:mysql:5.1.21:::::::*|cpe:2.3:a:oracle:mysql:5.1.21:::::::*|2019-12-17 15:06:50\n5.1.22|cpe:2.3:a:mysql:mysql:5.1.22:::::::*|cpe:2.3:a:oracle:mysql:5.1.22:::::::*|2019-12-17 15:06:56\n5.1.23:a|cpe:2.3:a:mysql:mysql:5.1.23:a::::::|cpe:2.3:a:oracle:mysql:5.1.23:a::::::|2019-12-17 15:23:28\n5.1.24|cpe:2.3:a:mysql:mysql:5.1.24:::::::*|cpe:2.3:a:oracle:mysql:5.1.24:::::::*|2019-12-17 15:07:03\n5.1.25|cpe:2.3:a:mysql:mysql:5.1.25:::::::*|cpe:2.3:a:oracle:mysql:5.1.25:::::::*|2019-12-17 15:07:10\n5.1.26|cpe:2.3:a:mysql:mysql:5.1.26:::::::*|cpe:2.3:a:oracle:mysql:5.1.26:::::::*|2019-12-17 15:07:17\n5.1.27|cpe:2.3:a:mysql:mysql:5.1.27:::::::*|cpe:2.3:a:oracle:mysql:5.1.27:::::::*|2019-12-17 15:07:24\n5.1.28|cpe:2.3:a:mysql:mysql:5.1.28:::::::*|cpe:2.3:a:oracle:mysql:5.1.28:::::::*|2019-12-17 15:07:31\n5.1.29|cpe:2.3:a:mysql:mysql:5.1.29:::::::*|cpe:2.3:a:oracle:mysql:5.1.29:::::::*|2019-12-17 15:07:38\n5.1.30|cpe:2.3:a:mysql:mysql:5.1.30:::::::*|cpe:2.3:a:oracle:mysql:5.1.30:::::::*|2019-12-17 15:07:46\n5.1.31:sp1|cpe:2.3:a:mysql:mysql:5.1.31:sp1::::::|cpe:2.3:a:oracle:mysql:5.1.31:sp1::::::|2019-12-17 15:23:35\n5.1.33|cpe:2.3:a:mysql:mysql:5.1.33:::::::*|cpe:2.3:a:oracle:mysql:5.1.33:::::::*|2019-12-17 15:07:55\n5.1.34:sp1|cpe:2.3:a:mysql:mysql:5.1.34:sp1::::::|cpe:2.3:a:oracle:mysql:5.1.34:sp1::::::|2019-12-17 15:23:42\n5.1.35|cpe:2.3:a:mysql:mysql:5.1.35:::::::*|cpe:2.3:a:oracle:mysql:5.1.35:::::::*|2019-12-17 15:08:03\n5.1.36|cpe:2.3:a:mysql:mysql:5.1.36:::::::*|cpe:2.3:a:oracle:mysql:5.1.36:::::::*|2019-12-17 15:08:10\n5.1.37:sp1|cpe:2.3:a:mysql:mysql:5.1.37:sp1::::::|cpe:2.3:a:oracle:mysql:5.1.37:sp1::::::|2019-12-17 15:23:49\n5.1.38|cpe:2.3:a:mysql:mysql:5.1.38:::::::*|cpe:2.3:a:oracle:mysql:5.1.38:::::::*|2019-12-17 15:08:17\n5.1.39|cpe:2.3:a:mysql:mysql:5.1.39:::::::*|cpe:2.3:a:oracle:mysql:5.1.39:::::::*|2019-12-17 10:22:31\n5.1.40|cpe:2.3:a:mysql:mysql:5.1.40:::::::*|cpe:2.3:a:oracle:mysql:5.1.40:::::::*|2019-12-17 10:22:46\n5.1.40:sp1|cpe:2.3:a:mysql:mysql:5.1.40:sp1::::::|cpe:2.3:a:oracle:mysql:5.1.40:sp1::::::|2019-12-17 10:22:39\n5.1.41|cpe:2.3:a:mysql:mysql:5.1.41:::::::*|cpe:2.3:a:oracle:mysql:5.1.41:::::::*|2019-12-17 10:22:52\n5.1.42|cpe:2.3:a:mysql:mysql:5.1.42:::::::*|cpe:2.3:a:oracle:mysql:5.1.42:::::::*|2019-12-17 10:22:59\n5.1.43|cpe:2.3:a:mysql:mysql:5.1.43:::::::*|cpe:2.3:a:oracle:mysql:5.1.43:::::::*|2019-12-17 10:23:06\n5.1.43:sp1|cpe:2.3:a:mysql:mysql:5.1.43:sp1::::::|cpe:2.3:a:oracle:mysql:5.1.43:sp1::::::|2019-12-17 10:23:11\n5.1.44|cpe:2.3:a:mysql:mysql:5.1.44:::::::*|cpe:2.3:a:oracle:mysql:5.1.44:::::::*|2019-12-17 10:23:17\n5.1.45|cpe:2.3:a:mysql:mysql:5.1.45:::::::*|cpe:2.3:a:oracle:mysql:5.1.45:::::::*|2019-12-17 10:23:24\n5.1.46|cpe:2.3:a:mysql:mysql:5.1.46:::::::*|cpe:2.3:a:oracle:mysql:5.1.46:::::::*|2019-12-17 10:23:38\n5.1.46:sp1|cpe:2.3:a:mysql:mysql:5.1.46:sp1::::::|cpe:2.3:a:oracle:mysql:5.1.46:sp1::::::|2019-12-17 10:23:31\n5.1.47|cpe:2.3:a:mysql:mysql:5.1.47:::::::*|cpe:2.3:a:oracle:mysql:5.1.47:::::::*|2019-12-17 10:23:45metadata{\:[\,\,\],\:\,\:true,\:true,\:[\,\,\],\:[\,\,\],\:{\:\},\:\}rubric{\:[\,\,\],\:[\,\,\],\:{\:\,\:\,\:\,\:[\,\,\,\],\:\,\:\,\:\,\:\,\:\}}all_involved_urlsnull |
| task_idnvd_003domainNVDautonomy_typeordered tableinstructionWe are cleaning up a batch of old SBOM matching rules that still represent Apache Tomcat 9.0.0 pre-release milestone versions using the legacy CPE form 9.0.0.Mx. We need to identify which milestone versions satisfy both of the following conditions: (1) they appear in the current NVD affected-product version ranges for the four historical vulnerabilities CVE-2016-5018, CVE-2016-6794, CVE-2016-6796, and CVE-2016-6797; and (2) the NVD change history shows that the CPE for these versions was remapped from the old short form (e.g., m1, m2) to the full form (e.g., milestone1, milestone2). Start from the Apache Tomcat 9.0.0 milestone versions, check each against the current NVD affected-product version ranges for these four CVEs, then examine the corresponding change history records. Retain only those milestone versions that satisfy both conditions. Exclude milestones that are included only because of other Tomcat 9.0.0 pre-release vulnerabilities, and do not include Tomcat 6, 7, 8, 8.5, or any non-Tomcat products. Finally, output an ordered multi-row table sorted by milestone number in ascending order, with columns: milestone | legacy_cpe | current_cpe | required_cves | remap_time_nvd.","start_url":"https://nvd.nist.gov/products/cpe/search","output_format":"Output as a Markdown multi-row table with fixed column names: milestone | legacy_cpe | current_cpe | required_cves | remap_time_nvd. Sort by milestone number in ascending order (M1, M2, ..., M9).","oracle_output_cardinality":9,"oracle_answer":"milestone|legacy_cpe|current_cpe|required_cves|remap_time_nvd\nM1|cpe:2.3:a:apache:tomcat:9.0.0:m1:::::::*|cpe:2.3:a:apache:tomcat:9.0.0:milestone1:::::::*|CVE-2016-5018;CVE-2016-6794;CVE-2016-6796;CVE-2016-6797|2023-12-08 11:41:18\nM2|cpe:2.3:a:apache:tomcat:9.0.0:m2:::::::*|cpe:2.3:a:apache:tomcat:9.0.0:milestone2:::::::*|CVE-2016-5018;CVE-2016-6794;CVE-2016-6796;CVE-2016-6797|2023-12-08 11:41:18\nM3|cpe:2.3:a:apache:tomcat:9.0.0:m3:::::::*|cpe:2.3:a:apache:tomcat:9.0.0:milestone3:::::::*|CVE-2016-5018;CVE-2016-6794;CVE-2016-6796;CVE-2016-6797|2023-12-08 11:41:18\nM4|cpe:2.3:a:apache:tomcat:9.0.0:m4:::::::*|cpe:2.3:a:apache:tomcat:9.0.0:milestone4:::::::*|CVE-2016-5018;CVE-2016-6794;CVE-2016-6796;CVE-2016-6797|2023-12-08 11:41:18\nM5|cpe:2.3:a:apache:tomcat:9.0.0:m5:::::::*|cpe:2.3:a:apache:tomcat:9.0.0:milestone5:::::::*|CVE-2016-5018;CVE-2016-6794;CVE-2016-6796;CVE-2016-6797|2023-12-08 11:41:18\nM6|cpe:2.3:a:apache:tomcat:9.0.0:m6:::::::*|cpe:2.3:a:apache:tomcat:9.0.0:milestone6:::::::*|CVE-2016-5018;CVE-2016-6794;CVE-2016-6796;CVE-2016-6797|2023-12-08 11:41:18\nM7|cpe:2.3:a:apache:tomcat:9.0.0:m7:::::::*|cpe:2.3:a:apache:tomcat:9.0.0:milestone7:::::::*|CVE-2016-5018;CVE-2016-6794;CVE-2016-6796;CVE-2016-6797|2023-12-08 11:41:18\nM8|cpe:2.3:a:apache:tomcat:9.0.0:m8:::::::*|cpe:2.3:a:apache:tomcat:9.0.0:milestone8:::::::*|CVE-2016-5018;CVE-2016-6794;CVE-2016-6796;CVE-2016-6797|2023-12-08 11:41:18\nM9|cpe:2.3:a:apache:tomcat:9.0.0:m9:::::::*|cpe:2.3:a:apache:tomcat:9.0.0:milestone9:::::::*|CVE-2016-5018;CVE-2016-6794;CVE-2016-6796;CVE-2016-6797|2023-12-08 11:41:18metadata{\:[\,\,\],\:\,\:true,\:true,\:[\,\,\],\:[\,\,\],\:{\:\},\:\}all_involved_urls{\:[\,\],\:[\,\,\,\],\:[\,\],\:[\,\,\,\,\,\,\,\,\],\:[\,\,\,\],\:[\,\,\],\:[\,\,\,\,\,\,\],\:[\,\]}rubric{\:[\,\,\],\:[\,\,\],\:{\:\,\:\,\:\,\:[\,\,\,\,\],\:\,\:\,\:\,\:\,\:\}} |
| task_idreptile_001domainREPTILE_DATABASEautonomy_typeordered tableoracle_output_cardinalityinstructionI am compiling a name concordance for early museum specimens of Indian snakes recorded in George A. Boulenger's 1890 work *The Fauna of British India, Including Ceylon and Burma*. Proceed as follows: first, using this specimen set as the scope, look up each corresponding current species page in the Reptile Database and retain only entries that simultaneously satisfy the following conditions: the species has since been reassigned to a different genus; the page still retains BMNH type specimen information; and the type locality still corresponds to India. For each qualifying entry, identify from the names listed on that page the oldest binomial that is closest to the present but has not yet been updated to the current genus name. Then, search again using this full old name and determine whether the search result points to only this single record. Finally, output the results as a single line sorted alphabetically by current accepted name, in the format <accepted name>|<old label>|<UNIQUE/AMBIG>|<BMNH status>; separate entries with semicolons. If no qualifying entries exist, output NONE.","start_url":"https://reptile-database.reptarium.cz/advanced_search","output_format":"Output the results as a single line sorted alphabetically by current accepted name, in the format <accepted name>|<old label>|<UNIQUE/AMBIG>|<BMNH status>; separate entries with semicolons. If no qualifying entries exist, output NONE.","oracle_answer":"Argyrophis oatesii|Typhlops oatesii|UNIQUE|SYNTYPES;Dendrelaphis grandoculis|Ahaetulla grandoculis|UNIQUE|LECTOTYPE;Fowlea sanctijohannis|Xenochrophis sanctijohannis|UNIQUE|HOLOTYPE;Gerrhopilus beddomii|Typhlops beddomii|UNIQUE|SYNTYPES;Hebius khasiensis|Amphiesma khasiense|UNIQUE|LECTOTYPE;Hebius parallelus|Amphiesma parallelum|AMBIG|LECTOTYPE;Indotyphlops jerdoni|Typhlops jerdoni|UNIQUE|SYNTYPES;Myriopholis blanfordii|Leptotyphlops blanfordi|AMBIG|SYNTYPES","metadata":"{\"State-Gated Retrieval\":[\"Candidates must come from the historical museum entries of Indian snakes in Boulenger 1890 *Fauna of British India*.\",\"The current Reptile Database species page must simultaneously satisfy: the species has been reassigned to a different genus; the page still retains BMNH type specimen information; the type locality still corresponds to India.\",\"On each qualifying page, select the oldest binomial that is closest to the present but has not yet been updated to the current genus name, then perform an exact back-search using this old name and label it UNIQUE or AMBIG.\"],\"dependency_type\":\"Data + Control\",\"intra_chain\":true,\"inter_chain\":true,\"data_dependency\":[\"the Boulenger 1890 historical sample and current accepted-species pages yield candidates that changed genus while retaining BMNH type data and India-compatible type locality\",\"each accepted page yields the nearest pre-current-genus old label, the BMNH type status, and the historical/type-locality evidence anchoring the entity\",\"final rows are produced only after the old-label exact back-search is evaluated for uniqueness without losing the original entity anchor\"],\"control_dependency\":[\"when an exact old-label back-search now hits multiple species, the original anchored entity must be kept and the uniqueness label must be set to AMBIG\",\"once a candidate has passed the historical-cluster, genus-shift, BMNH, and India-locality gates, later old-label ambiguity can only rewrite the uniqueness field, not the entity itself\",\"candidate anchoring must be preserved across the two-stage workflow so that old-label search noise does not redirect the accepted-species row\"],\"freeze\":{\"historical_window\":\"the George A. Boulenger 1890 British India sample, evaluated against current Reptile Database accepted pages, synonym labels, BMNH type data, and present type-locality text\"},\"answer_type\":\"multi-row ordered table\"}","rubric":"{\"inclusion_conditions\":[\"Candidates must come from the historical museum entries of Indian snakes in Boulenger 1890 *Fauna of British India*.\",\"The current Reptile Database species page must simultaneously satisfy: the species has been reassigned to a different genus; the page still retains BMNH type specimen information; the type locality still corresponds to India.\",\"On each qualifying page, select the oldest binomial that is closest to the present but has not yet been updated to the current genus name, then perform an exact back-search using this old name and label it UNIQUE or AMBIG.\"],\"exclusion_conditions\":[\"Exclude entries where the current accepted name has not undergone a genus change, or where the page no longer retains BMNH type specimen information.\",\"Exclude entries whose type locality does not correspond to India.\",\"Do not discard candidates whose exact old-name back-search is non-unique; such candidates should be retained and labeled AMBIG.\"],\"normalization\":{\"field_separator\":\"|\",\"record_separator\":\";\",\"schema\":[\"accepted name\",\"old label\",\"UNIQUE/AMBIG\",\"BMNH status\"],\"accepted_name\":\"emit the current accepted binomial exactly as shown in the reconciled source\",\"old_label\":\"preserve the historical label exactly as used on the legacy branch\",\"UNIQUE/AMBIG\":\"emit `UNIQUE` or `AMBIG` in uppercase\",\"BMNH_status\":\"normalize the type-status field to uppercase labels such as HOLOTYPE, LECTOTYPE, or SYNTYPES\",\"sorting_or_selection\":\"accepted name alphabetical ascending\"}}","all_involved_urls":"null"} |
| {"task_id":"reptile_002","domain":"REPTILE_DATABASE","autonomy_type":"ordered table","oracle_output_cardinality":3,"instruction":"I am compiling current name correspondences for early museum specimens of Southeast Asian Lygosoma. The old labels correspond to Riopa combinations attributed to Smith (1935/1937) in the Synonym section of each species page, but some of these old names now map to multiple current species under the current classification. Process as follows: first, within the current Lygosoma checklist, examine each current species page and locate these Smith (1935/1937) Riopa combinations in the Synonym section; then compare these old names, identify those that point to multiple current species, and list all involved current species. Next, for each qualifying current species, record the first-listed type specimen number, type category, and the region ultimately corresponding to the type locality on that page. Finally, output the results as a single line with entries sorted alphabetically by current accepted name, in the format <accepted>|<Smith-era Riopa label>|<first type code>|<type status>|<region>; separate entries with semicolons; if no qualifying entries exist, output NONE.","start_url":"https://reptile-database.reptarium.cz/advanced_search","output_format":"Output the results as a single line with entries sorted alphabetically by current accepted name, in the format <accepted>|<Smith-era Riopa label>|<first type code>|<type status>|<region>; separate entries with semicolons; if no qualifying entries exist, output NONE.","oracle_answer":"Lygosoma bampfyldei|Riopa bampfyldei|BMNH 1946.8.10.84|SYNTYPES|East Malaysia;Lygosoma peninsulare|Riopa bampfyldei|LSUHC 13857|HOLOTYPE|Peninsular Malaysia;Lygosoma schneideri|Riopa bampfyldei|NMBA 4743|HOLOTYPE|Indonesia","metadata":"{\"State-Gated Retrieval\":[\"Retain only those accepted species that can be traced from the Smith 1937 old-name chain to a current Lygosoma record and for which the type accession, type status, and locality can be read on the current page.\",\"For each accepted record, output the accepted name, the corresponding old label, the type accession, the type status, and the locality.\",\"If the same exact old name is attached to multiple accepted species, retain them as separate chains.\"],\"dependency_type\":\"Data + Control\",\"intra_chain\":true,\"inter_chain\":true,\"data_dependency\":[\"the Smith 1937 old-name chain yields the three current Lygosoma accepted pages connected to Riopa bampfyldei / Lygosoma bampfyldei\",\"each accepted page yields its own type accession, type status, and type locality needed to separate the homonymous old-name paths\",\"final rows are produced only after the one-to-many old-name split is rebuilt and the three paths are re-anchored to their own type evidence\"],\"control_dependency\":[\"same-specific-epithet matching cannot establish a one-to-one mapping when the same Smith 1937 old name appears on multiple accepted pages\",\"once literal old-name reuse is exposed, the workflow must keep the original parallel three-chain state rather than collapsing to a one-to-one mapping\",\"type accession and locality evidence must be read separately for each chain rather than patched onto the first stopping page\"],\"freeze\":{\"historical_window\":\"the current Reptile Database accepted pages linked to the Smith 1937 old-name cluster around Riopa bampfyldei, using current type accessions, statuses, and localities\"},\"answer_type\":\"multi-row ordered table\"}","rubric":"{\"inclusion_conditions\":[\"Retain only those accepted species that can be traced from the Smith 1937 old-name chain to a current Lygosoma record and for which the type accession, type status, and locality can be read on the current page.\",\"For each accepted record, output the accepted name, the corresponding old label, the type accession, the type status, and the locality.\",\"If the same exact old name is attached to multiple accepted species, retain them as separate chains.\"],\"exclusion_conditions\":[\"Exclude results that assume the same species based solely on the old name and current specific epithet being identical.\",\"Exclude results that retain only the first chain after discovering the same exact old name on multiple accepted pages.\",\"Exclude results that do not return to each accepted page to re-verify the type accession and locality.\"],\"normalization\":{\"field_separator\":\"|\",\"record_separator\":\";\",\"schema\":[\"accepted\",\"Smith-era Riopa label\",\"first type code\",\"type status\",\"region\"],\"accepted\":\"emit the current accepted binomial exactly as shown in the reconciled source\",\"Smith-era_Riopa_label\":\"preserve the Smith-era Riopa label exactly as used on the legacy branch\",\"first_type_code\":\"emit the first accession code that follows the selected type-status statement\",\"type_status\":\"normalize the type-status field to uppercase labels such as HOLOTYPE or SYNTYPES\",\"region\":\"emit the modern geographic region label from the source reconciliation, such as East Malaysia or Peninsular Malaysia\",\"sorting_or_selection\":\"accepted name alphabetical ascending\"}}","all_involved_urls":"null"} |
| {"task_id":"reptile_003","domain":"REPTILE_DATABASE","autonomy_type":"ordered table","oracle_output_cardinality":10,"instruction":"I am organizing current-name cross-references for a batch of early museum catalog cards that were still filed under Mabuya or Mabuia around 1980, and need to map them into the current Eutropis checklist. Process as follows: first, examine each species page in the current Eutropis checklist one by one, and on each page identify the first synonym label that is both in Mabuya or Mabuia form and tagged with the 1980 citation year. Then, perform an exact name search using that 1980 old label, and retain only those species for which the back-search result points to more than one current Eutropis entry. For each qualifying entry, record its current accepted name, the 1980 old label, the qualifier annotation for that old label on the page, and the first type category and code listed on the current page. Finally, output a single line per entry in the format <accepted>|<1980 label>|<qualifier>|<type status>|<first type code>, sorted alphabetically by old label and, within the same old label, by accepted name. Join entries with semicolons. If no qualifying entries exist, output NONE.","start_url":"https://reptile-database.reptarium.cz/advanced_search","output_format":"Output a single line per entry in the format <accepted>|<1980 label>|<qualifier>|<type status>|<first type code>, sorted alphabetically by old label and, within the same old label, by accepted name. Join entries with semicolons. If no qualifying entries exist, output NONE.","oracle_answer":"Eutropis indeprensa|Mabuya indeprensa|NONE|HOLOTYPE|CAS 85664;Eutropis sahulinghangganan|Mabuya indeprensa|PART-MISID|HOLOTYPE|PNM 9867;Eutropis multifasciata|Mabuya multifasciata|NONE|NEOTYPE|MZB 11912;Eutropis multicarinata|Mabuya multicarinata|NONE|HOLOTYPE|BMNH 1946.8.15.13-15;Eutropis borealis|Mabuya multicarinata borealis|NONE|HOLOTYPE|CAS 15447;Eutropis islamaliit|Mabuya multicarinata borealis|PART-MISID|HOLOTYPE|PNM 9847;Eutropis caraga|Mabuya multicarinata multicarinata|PART-MISID|HOLOTYPE|PNM 9845;Eutropis palauensis|Mabuya multicarinata multicarinata|PART-BYIMPL|HOLOTYPE|CAS 248251;Eutropis dattaroyi|Mabuya rudis|PART|HOLOTYPE|ZSI 25118;Eutropis rudis|Mabuya rudis|NONE|LECTOTYPE|BMNH 1946.8.15.26","metadata":"{\"State-Gated Retrieval\":[\"Retain only those current Eutropis accepted species for which the synonym block contains, scanning top-down, a first synonym label that begins with Mabuya or Mabuia and has the year 1980.\",\"Then, perform an exact name search using that 1980 old label, and output the accepted name, old label, back-search qualifier, type status, and type specimen.\",\"If the subsequent back-search reveals additional matching entries, revisit the earlier results and include them in the final output.\"],\"dependency_type\":\"Data + Control\",\"intra_chain\":true,\"inter_chain\":true,\"data_dependency\":[\"current Eutropis detail pages yield the first 1980-tagged Mabuya/Mabuia synonym label for each accepted species\",\"exact back-search on that 1980 old label yields the present accepted page, type status, and type specimen needed for the output row\",\"final rows are produced only after the 1980-specific old-label anchor is rebuilt and the accepted-species set is expanded when back-search evidence reveals omitted entries\"],\"control_dependency\":[\"the workflow must identify the first Mabuya/Mabuia label carrying the 1980 year rather than stopping at the first old label on the page\",\"when old-label back-search reveals missing entries such as Eutropis dattaroyi, the accepted-species set must be expanded to include them rather than patched in place\",\"accepted-page synonym parsing and back-search interpretation must stay coupled so that 1980 labels are not replaced by older but irrelevant names\"],\"freeze\":{\"historical_window\":\"current Reptile Database Eutropis pages, restricted to the first synonym label that is both Mabuya/Mabuia-form and tagged with the 1980 citation year\"},\"answer_type\":\"multi-row ordered table\"}","rubric":"{\"inclusion_conditions\":[\"Retain only those current Eutropis accepted species for which the synonym block contains, scanning top-down, a first synonym label that begins with Mabuya or Mabuia and has the year 1980.\",\"Then, perform an exact name search using that 1980 old label, and output the accepted name, old label, back-search qualifier, type status, and type specimen.\",\"If the subsequent back-search reveals additional matching entries, revisit the earlier results and include them in the final output.\"],\"exclusion_conditions\":[\"Exclude results that mistakenly treat the first old name on the page as the first 1980 old label.\",\"Exclude results whose exact back-search on the 1980 old label points to only a single current Eutropis entry.\",\"Exclude results that, when subsequent evidence reveals omitted entries, do not revisit the earlier results and rebuild the accepted-species set.\"],\"normalization\":{\"field_separator\":\"|\",\"record_separator\":\";\",\"schema\":[\"accepted\",\"1980 label\",\"qualifier\",\"type status\",\"first type code\"],\"accepted\":\"emit the current accepted binomial exactly as shown in the reconciled source\",\"1980_label\":\"preserve the 1980-era label exactly as used on the legacy branch\",\"qualifier\":\"normalize to one of `NONE`, `PART`, `PART-MISID`, or `PART-BYIMPL`\",\"type_status\":\"normalize the type-status field to uppercase labels such as HOLOTYPE or SYNTYPES\",\"first_type_code\":\"emit the first accession code that follows the selected type-status statement\",\"sorting_or_selection\":{\"primary\":\"1980 label alphabetical ascending\",\"secondary\":\"accepted name alphabetical ascending\"}}}","all_involved_urls":"null"} |
| {"task_id":"reptile_004","domain":"REPTILE_DATABASE","autonomy_type":"ordered table","oracle_output_cardinality":6,"instruction":"I am organizing a batch of early museum catalog cards for half-toed geckos that still use the old labels listed in the comments section of the Hemidactylus brookii species page. Process as follows: first examine the old labels listed in the comments section of the current Hemidactylus brookii species page, and select those that are no longer placed in Hemidactylus brookii and for which the database has clearly provided their current assignment. Then, for each qualifying label, determine the current species it should ultimately be assigned to; if merging from that label to the current species must pass through an intermediate old name, write that intermediate name, otherwise write DIRECT. At the same time, record the type status, type code, and the country of the type locality for that label itself. Finally, output a single line with entries sorted alphabetically by legacy label: <legacy label>|<accepted>|<bridge>|<type status>|<type code>|<country>; separate entries with semicolons; if no qualifying entries exist, output NONE.","start_url":"https://reptile-database.reptarium.cz/advanced_search","output_format":"Output a single line with entries sorted alphabetically by legacy label: <legacy label>|<accepted>|<bridge>|<type status>|<type code>|<country>; separate entries with semicolons; if no qualifying entries exist, output NONE.","oracle_answer":"Hemidactylus brooki leightoni|Hemidactylus angulatus|DIRECT|HOLOTYPE|BMNH 1946.8.25.65|Colombia;Hemidactylus brookii haitianus|Hemidactylus angulatus|DIRECT|LECTOTYPE|ZMH R22378|Haiti;Hemidactylus brookii parvimaculatus|Hemidactylus parvimaculatus|DIRECT|HOLOTYPE|NMSL No. RG. 15|Sri Lanka;Hemidactylus brookii subtriedroides|Hemidactylus murrayi|Hemidactylus tenkatei|LECTOTYPE|BMNH [subtriedroides] [tenkatei]|Myanmar;Hemidactylus gleadowi|Hemidactylus gleadowi|DIRECT|NEOTYPE|BMNH 1884.7.25.8|Pakistan;Hemidactylus tenkatei|Hemidactylus murrayi|DIRECT|LECTOTYPE|RMNH 4353|Indonesia","metadata":"{\"State-Gated Retrieval\":[\"Each result must start from the old-label chain specified in the task and ultimately determine the accepted name, bridge status, type status, type accession, and locality from the current accepted page.\",\"For the subtriedroides branch, continue along tenkatei to murrayi, output the final accepted name as Hemidactylus murrayi, and retain bridge=Hemidactylus tenkatei.\",\"For branches such as haitianus and leightoni, read their own locality/type lines rather than borrowing header information from the wrong page.\"],\"dependency_type\":\"Data + Control\",\"intra_chain\":true,\"inter_chain\":true,\"data_dependency\":[\"The old-label chains from Hemidactylus angulatus and related names yield accepted pages, bridge names, and the type/locality lines for each path.\",\"Comment, Types, and type-comment blocks yield the later evidence that tenkatei is further sunk into H. murrayi.\",\"Final rows are produced only after each old-label path is re-anchored to its own locality/type line and the subtriedroides -> tenkatei -> murrayi bridge chain is completed.\"],\"control_dependency\":[\"Type data must be taken from the locality/type line belonging to the actual old-label branch rather than from the header block of the wrong accepted page.\",\"Stopping at the intermediate tenkatei bridge is insufficient; the workflow must continue to the later murrayi evidence and rewrite both tenkatei and subtriedroides final accepted names.\",\"Bridge labels remain useful, but final accepted names and type lines must be determined from the deeper synonym chain.\"],\"freeze\":{\"historical_window\":\"Current Reptile Database pages for the Hemidactylus angulatus / brookii / murrayi complex, using present locality, type, comment, and type-comment blocks.\"},\"answer_type\":\"multi-row ordered table\"}","rubric":"{\"inclusion_conditions\":[\"Each result must start from the old-label chain specified in the task and ultimately determine the accepted name, bridge status, type status, type accession, and locality from the current accepted page.\",\"For the subtriedroides branch, continue along tenkatei to murrayi, output the final accepted name as Hemidactylus murrayi, and retain bridge=Hemidactylus tenkatei.\",\"For branches such as haitianus and leightoni, read their own locality/type lines rather than borrowing header information from the wrong page.\"],\"exclusion_conditions\":[\"Exclude results that follow the wrong page hierarchy and apply the Holotype/locality from the H. angulatus page header to other branches.\",\"Exclude results that stop early on the subtriedroides -> tenkatei chain without continuing to read the murrayi evidence.\",\"Exclude results that fail to align the locality/type line with each specific old-label branch.\"],\"normalization\":{\"field_separator\":\"|\",\"record_separator\":\";\",\"schema\":[\"legacy label\",\"accepted\",\"bridge\",\"type status\",\"type code\",\"country\"],\"legacy_label\":\"Preserve the legacy label exactly as used on the source branch.\",\"accepted\":\"Emit the current accepted binomial exactly as shown in the reconciled source.\",\"bridge\":\"Emit `DIRECT` when no intermediate legacy bridge name is needed; otherwise emit the intermediate legacy name that links the branches.\",\"type_status\":\"Normalize the type-status field to uppercase labels such as HOLOTYPE, LECTOTYPE, or NEOTYPE.\",\"type_code\":\"Take the type accession from the legacy-label branch rather than from a generic type line elsewhere on the page.\",\"country\":\"Emit the modern English country name from the reconciled source branch.\",\"sorting_or_selection\":\"Legacy label alphabetical ascending.\"}}","all_involved_urls":"null"} |
| {"task_id":"reptile_005","domain":"REPTILE_DATABASE","autonomy_type":"ordered table","oracle_output_cardinality":2,"instruction":"I am checking a synonymization record in the current Naja species list that involves an early Sri Lankan museum name, and I want to organize the key information from the accepted side and the contested-name side into a two-row comparison table. First, in the current Naja list, locate the contested-name record associated with Sri Lanka. Then verify its recent-use evidence, the database's final treatment, the correction concerning lectotype assignment, the restricted type locality, and the holotype information for the contested name itself. Finally, output a two-row comparison table with columns rank|name_role|name|status_or_evidence|specimen_role|specimen_code|supporting_note. Row 1 should be ACCEPTED_NAME and give the accepted name, final status, valid lectotype, and restricted type locality. Row 2 should be CONTESTED_NAME and give the contested name, recent-use evidence, holotype, and the corrected invalid lectotype claim.","start_url":"https://reptile-database.reptarium.cz/advanced_search?genus=Naja&submit=Search","output_format":"Output a two-row comparison table with columns rank|name_role|name|status_or_evidence|specimen_role|specimen_code|supporting_note.","oracle_answer":"1|ACCEPTED_NAME|Naja naja|SUBJECTIVE_JUNIOR_SYNONYM_OF_NAJA_NAJA|VALID_LECTOTYPE|ZMB 2796|restricted type locality: Sri Lanka\n2|CONTESTED_NAME|Naja polyocellata|SHI et al. 2022+SILVA et al. 2023|HOLOTYPE|BMNH 1946.1.18.50|invalid lectotype claim: NRM (= NHR) Lin-90metadata{\:[\,\,\],\:\,\:true,\:true,\:[\,\,\],\:[\,\,\],\:{\:\},\:\}rubric{\:[\,\,\],\:[\,\,\],\:{\:\,\:\,\:\,\:[\,\,\,\,\,\,\],\:\,\:[\,\],\:\,\:\,\:\}}all_involved_urlsnull |
| task_idwateroffice_001domainWATER_OFFICEautonomy_typeordered tableoracle_output_cardinalityinstructionI am compiling historical daily flow comparison data for the Saint John River spring freshet period and want to determine, along the mainstem, which station in the current monitoring network is the last one that still allows daily flow comparisons for March–April 1973 and March–April 2018, and which mainstem station downstream of it is the first one that no longer satisfies that condition. First, use the current monitoring network shown on the Wateroffice real-time page as of April 16, 2026, and retain only stations on the Saint John River mainstem. Then, for each station, check the historical availability page, distinguish Flow Data Availability from Daily Water Level Data Availability, and determine whether March–April 1973 and March–April 2018 both have complete daily flow records. If the target windows satisfy the complete daily flow condition, set target_window_status to PASS and final_verdict to KEEP. If the target years have only water level records and no daily flow records, set target_window_status to LEVEL_ONLY and final_verdict to REJECT. If daily flow records exist but any target month in March–April 1973 or March–April 2018 is missing or incomplete, set target_window_status to FLOW_GAP and final_verdict to REJECT. Finally, output a two-row breakpoint table with columns rank|station_number|breakpoint_role|target_window_status|final_verdict. In the first row, give the last usable station and set breakpoint_role to LAST_USABLE. In the second row, give the first failing station encountered downstream and set breakpoint_role to FIRST_DOWNSTREAM_FAIL.","start_url":"https://wateroffice.ec.gc.ca/search/real_time_e.html","output_format":"Output a two-row breakpoint table with columns rank|station_number|breakpoint_role|target_window_status|final_verdict.","oracle_answer":"1|01AF002|LAST_USABLE|PASS|KEEP\n2|01AK003|FIRST_DOWNSTREAM_FAIL|LEVEL_ONLY|REJECTmetadata{\:[\,\,\],\:\,\:true,\:true,\:[\,\,\],\:[\,\,\],\:{\:\},\:\}rubric{\:[\,\,\],\:[\,\,\],\:{\:\,\:\,\:\,\:[\,\,\,\,\],\:\,\:\,\:\,\:\,\:\,\:[\,\]}}all_involved_urlsnull |
| task_idwateroffice_002domainWATER_OFFICEautonomy_typeordered tableoracle_output_cardinalityinstructionI am compiling historical daily flow records for the Bulkley River reach from Houston to Hazelton and want to verify, station by station, whether candidate mainstem stations on the Bulkley River truly have complete daily flow records for April–July 1948 and April–July 1950. First, from the Historical Results list, select candidate stations that lie on the Bulkley River mainstem, fall within the Houston-to-Hazelton reach, appear to cover both 1948 and 1950, and carry a Flow label. Then open each station's availability page and check whether the target years fall within the Flow section and whether April-July is marked C throughout. If the target windows satisfy the complete daily flow condition, set target_window_status to PASS and final_verdict to KEEP. If the target years have only water level records and no daily flow records, set target_window_status to LEVEL_ONLY and final_verdict to REJECT. If daily flow records exist but any month from April to July is not complete, set target_window_status to FLOW_GAP and final_verdict to REJECT. Finally, output a table in downstream-to-upstream order with columns rank|station_number|target_window_status|final_verdict.","start_url":"https://wateroffice.ec.gc.ca/search/historical_e.html","output_format":"Output a table in downstream-to-upstream order with columns rank|station_number|target_window_status|final_verdict.","oracle_answer":"1|08EE004|PASS|KEEP\n2|08EE001|LEVEL_ONLY|REJECT\n3|08EE005|FLOW_GAP|REJECT\n4|08EE003|FLOW_GAP|REJECT","metadata":"{\"State-Gated Retrieval\":[\"First identify mainstem candidate stations on the Historical Results list that appear to cover 1948/1950 and carry a Flow label.\",\"A station qualifies only if its availability page shows the target years in the Flow section and all of April-July marked C.\",\"The final output must report target_window_status and final_verdict for each station in the same station set, ordered from downstream to upstream.\"],\"dependency_type\":\"Data + Control\",\"intra_chain\":true,\"inter_chain\":true,\"data_dependency\":[\"the Historical Results list yields four downstream-to-upstream mainstem candidates whose result rows span the target years and include Flow labels\",\"each candidate's availability page yields whether the target years are in the Flow table and whether April-July is complete\the final ordered rows are produced only after all four stations are classified under the full Flow-table-and-Apr-Jul-C rule\control_dependency\list-level year spans and coarse Flow tags must be confirmed by explicit target-month checks on the availability page\Hazelton, Smithers, and Houston candidate statuses must be fixed before the downstream-most KEEP row can be confirmed\row order and KEEP/REJECT labels must be emitted only after the full availability-page rule is applied to all four stations\freeze\historical_window\the target mainstem stations tested against 1948 and 1950 daily Flow availability, with April-July requiring complete C coverage\answer_type\multi-row ordered table\,:inclusion_conditions\First, from the Historical Results list, identify mainstem candidate stations that superficially appear to cover 1948/1950 and have a Flow label.\A truly qualified station must satisfy, on its availability page, that the target years are in the Flow section and that April–July are all marked C.\Finally, report target_window_status and final_verdict for each station in the same station set, ordered from downstream to upstream.\exclusion_conditions\Exclude stations deemed usable based solely on list-level year span and coarse Flow label.\Exclude retaining LEVEL_ONLY or FLOW_GAP stations in the final answer.\Exclude results that do not apply the full availability-page rule to all initially screened stations.\normalization\field_separator\|\record_separator\\\n\header_row\omit the header row; emit data rows only\schema\rank\station_number\target_window_status\final_verdict\rank\emit 1 through 4 in downstream-to-upstream order within the validated station set\station_number\preserve the station number exactly\target_window_status\emit `PASS`, `LEVEL_ONLY`, or `FLOW_GAP` exactly\final_verdict\emit `KEEP` or `REJECT` exactly\sorting_or_selection\downstream to upstream within the validated Bulkley River mainstem station set\stop_condition\all four initially screened stations are classified under the full availability-page rule\the ordered table is emitted only after the downstream-to-upstream sequence is fixed\,:} |
| {:,:,:,:5,:,:,:,:,:State-Gated Retrieval\All five candidate stations must be verified on the availability page to confirm that the target years are present in the Flow table and that May–July is entirely marked C.\Each station must output a PASS / LEVEL_ONLY / FLOW_GAP judgment for both target years, and then be marked KEEP or REJECT accordingly.\The result order must follow the river-path ordering specified in the task.\dependency_type\Data + Control\intra_chain\inter_chain\data_dependency\the river-path ordering yields five South Thompson / Thompson mainstem candidates\each station's availability page yields whether 1958 and 1971 are present in the Flow table and whether May-July is complete\",\"final rows are produced only after every candidate is classified under the stricter target-year and target-month Flow rule\"],\"control_dependency\":[\"overall year coverage on the results list must be confirmed by the availability page inside the Flow table\",\"Monte Creek, Savona, and Kamloops must be rejected by LEVEL_ONLY / FLOW_GAP evidence before the KEEP rows are final\",\"the river-order sequence remains fixed, but station statuses must be assigned only after the availability-page rule is applied to all five stations\"],\"freeze\":{\"historical_window\":\"the ordered South Thompson / Thompson mainstem candidates tested against 1958 and 1971 Flow availability, with May-July requiring complete C coverage\"},\"answer_type\":\"multi-row ordered table\"}","rubric":"{\"inclusion_conditions\":[\"All five candidate stations must be verified on the availability page to confirm that the target years are present in the Flow table and that May–July is entirely marked C.\",\"Each station must output a PASS / LEVEL_ONLY / FLOW_GAP judgment for both target years, and then be marked KEEP or REJECT accordingly.\",\"The result order must follow the river-path ordering specified in the task.\"],\"exclusion_conditions\":[\"Exclude results that treat overall year coverage on the list as sufficient evidence.\",\"Exclude results that fail to distinguish between Flow table target-year and month completeness versus Level-only evidence.\",\"Exclude results that do not apply the same availability-page rule uniformly to all five stations.\"],\"normalization\":{\"field_separator\":\"|\",\"record_separator\":\"\\n\",\"header_row\":\"omit the header row; emit data rows only\",\"schema\":[\"rank\",\"station_number\",\"station_name\",\"1958_MayJul\",\"1971_MayJul\",\"final_verdict\"],\"rank\":\"use the fixed Chase-to-Spences-Bridge mainstem order from the reference index, not the transient search-result order\",\"year_window_fields\":\"the two year-window fields emit only PASS, LEVEL_ONLY, or FLOW_GAP\",\"final_verdict\":\"emit `KEEP` or `REJECT` exactly\"}}","all_involved_urls":"null"} |
| {"task_id":"wateroffice_004","domain":"WATER_OFFICE","autonomy_type":"ordered table","oracle_output_cardinality":13,"instruction":"I am compiling historical daily flow records for the mainstem Columbia River from Fairmont Hot Springs to the International Boundary. I want to check, station by station in Wateroffice, which stations have complete daily flow records for May–October of 1955 and 1956. First, locate stations on the Columbia River mainstem between Fairmont Hot Springs and the International Boundary whose records span both 1955 and 1956. Then, proceeding from upstream to downstream, check the availability of daily flow for May–October of 1955 and 1956 at each station. If a station has only water-level records and no daily flow records for a given target year, mark that year as LEVEL_ONLY. If daily flow for May–October of that year is complete, mark it as PASS. Otherwise, if the year has no usable daily flow records or any month within May–October is incomplete, mark it as FLOW_GAP. Only when both years satisfy the PASS condition should the final_verdict be KEEP; otherwise write REJECT. Finally, output a table in upstream-to-downstream order with columns: rank|station_number|station_name|1955_MayOct|1956_MayOct|final_verdict.","start_url":"https://wateroffice.ec.gc.ca/search/historical_e.html","output_format":"Output a table in upstream-to-downstream order with columns: rank|station_number|station_name|1955_MayOct|1956_MayOct|final_verdict.","oracle_answer":"1|08NA045|COLUMBIA RIVER NEAR FAIRMONT HOT SPRINGS|PASS|PASS|KEEP\n2|08NA004|COLUMBIA RIVER AT ATHALMER|LEVEL_ONLY|LEVEL_ONLY|REJECT\n3|08NA052|COLUMBIA RIVER NEAR EDGEWATER|PASS|FLOW_GAP|REJECT\n4|08NA002|COLUMBIA RIVER AT NICHOLSON|PASS|PASS|KEEP\n5|08NB005|COLUMBIA RIVER AT DONALD|PASS|PASS|KEEP\n6|08NB006|COLUMBIA RIVER AT SURPRISE RAPIDS|PASS|PASS|KEEP\n7|08ND007|COLUMBIA RIVER ABOVE NAGLE CREEK|PASS|PASS|KEEP\n8|08ND011|COLUMBIA RIVER ABOVE STEAMBOAT RAPIDS|FLOW_GAP|PASS|REJECT\n9|08ND006|COLUMBIA RIVER AT TWELVE MILE FERRY|PASS|PASS|KEEP\n10|08NE002|COLUMBIA RIVER AT CASTLEGAR|LEVEL_ONLY|LEVEL_ONLY|REJECT\n11|08NE049|COLUMBIA RIVER AT BIRCHBANK|PASS|PASS|KEEP\n12|08NE003|COLUMBIA RIVER AT TRAIL|LEVEL_ONLY|LEVEL_ONLY|REJECT\n13|08NE058|COLUMBIA RIVER AT INTERNATIONAL BOUNDARY|PASS|PASS|KEEP","metadata":"{\"State-Gated Retrieval\":[\"Every candidate Columbia River station must be checked on its availability page to confirm that the target years appear in the Flow table and that May–October is fully marked C.\",\"For each station, provide the KEEP/REJECT basis for both 1955 and 1956, distinguishing LEVEL_ONLY and FLOW_GAP cases from stations that only appear eligible from their overall year span.\",\"The final table must preserve the order and fields specified in the task.\"],\"dependency_type\":\"Data + Control\",\"intra_chain\":true,\"inter_chain\":true,\"data_dependency\":[\"the Columbia River candidate ordering yields fourteen stations spanning the target years\",\"each availability page yields whether 1955 and 1956 are present in the Flow table and whether May-October is complete\",\"final rows are produced only after every station is classified under the stricter Flow-table-and-May-Oct-C rule\"],\"control_dependency\":[\"list-level year spans must be confirmed by the availability page inside the Flow table\",\"LEVEL_ONLY and FLOW_GAP evidence must be applied to each station before the KEEP/REJECT table is finalized\",\"station-by-station verdicts and rejection reasons must be determined only after the full availability-page rule replaces the initial year-span screen\"],\"freeze\":{\"historical_window\":\"the ordered Columbia River candidates tested against 1955 and 1956 Flow availability, with May-October requiring complete C coverage\"},\"answer_type\":\"multi-row ordered table\"}","rubric":"{\"inclusion_conditions\":[\"Every candidate Columbia River station must be checked on its availability page to confirm that the target years appear in the Flow table and that May–October is fully marked C.\",\"For each station, provide the KEEP/REJECT basis for both 1955 and 1956, distinguishing LEVEL_ONLY and FLOW_GAP cases from stations that only appear eligible from their overall year span.\",\"The final table must preserve the order and fields specified in the task.\"],\"exclusion_conditions\":[\"Exclude results that retain a station solely because its overall record spans 1955/1956.\",\"Exclude results that fail to distinguish, station by station, between Flow availability, Level-only, and incomplete months.\",\"Exclude results that do not apply the full availability-page rule uniformly to all 14 stations.\"],\"normalization\":{\"field_separator\":\"|\",\"record_separator\":\"\\n\",\"header_row\":\"omit the header row; emit data rows only\",\"schema\":[\"rank\",\"station_number\",\"station_name\",\"1955_MayOct\",\"1956_MayOct\",\"final_verdict\"],\"rank\":\"use the fixed upstream-to-downstream Columbia mainstem order from the reference index, not the transient search-result order\",\"year_window_fields\":\"the two year-window fields emit only PASS, LEVEL_ONLY, or FLOW_GAP\",\"final_verdict\":\"emit `KEEP` or `REJECT` exactly\"}}","all_involved_urls":"null"} |
| {"task_id":"wateroffice_005","domain":"WATER_OFFICE","autonomy_type":"ordered table","oracle_output_cardinality":3,"instruction":"I am organizing historical daily-flow records for the Peace River reach from Dunvegan Bridge to Fifth Meridian and want to verify three things at once: the true furthest-downstream reusable station on the mainstem, the first failing station encountered after it, and the provisional terminal station that a station-level screen based only on year coverage and record type would mistakenly treat as the downstream winner. Screen only among Peace River mainstem stations between Dunvegan Bridge and Fifth Meridian. First, from the station list, identify provisional candidates that appear to cover both 1966 and 1967 and show a daily-flow record type. Then open each availability page and verify whether May through September in both 1966 and 1967 appears in the Flow table and is complete for every month. If both target years satisfy the complete daily-flow condition, write PASS in target_window_status. If a target year has only level records and no daily-flow records, write LEVEL_ONLY. If daily-flow records exist but any target month from May through September is incomplete, write FLOW_GAP. If the target year has neither daily-flow nor level records, write NO_TARGET_RECORD. Finally, output a three-row diagnostic table with columns rank|station_number|diagnosis_role|target_window_status|provisional_count|final_count. The rows must be FINAL_STATION, FIRST_BREAK_AFTER_FINAL, and PROVISIONAL_STATION in that order, and the last two columns must repeat the global provisional_count and final_count in every row.","start_url":"https://wateroffice.ec.gc.ca/search/historical_e.html","output_format":"Output a three-row diagnostic table with columns rank|station_number|diagnosis_role|target_window_status|provisional_count|final_count.","oracle_answer":"1|07HA001|FINAL_STATION|PASS|5|2\n2|07HD001|FIRST_BREAK_AFTER_FINAL|FLOW_GAP|5|2\n3|07KA002|PROVISIONAL_STATION|FLOW_GAP|5|2","metadata":"{\"State-Gated Retrieval\":[\"Screen only among Peace River mainstem stations between Dunvegan Bridge and Fifth Meridian.\",\"A genuinely reusable station must satisfy the availability-page rule that both 1966 and 1967 appear in the Flow table and that every month from May through September is complete.\",\"The final output must jointly report FINAL_STATION, FIRST_BREAK_AFTER_FINAL, and PROVISIONAL_STATION, while repeating provisional_count and final_count in every row.\"],\"dependency_type\":\"Data + Control\",\"intra_chain\":true,\"inter_chain\":true,\"data_dependency\":[\"the Peace River mainstem chain between Dunvegan Bridge and Fifth Meridian yields the initial station list and the downstream breakpoint positions that need revalidation\",\"each availability page yields whether 1966 and 1967 are present in the Flow or Level tables and whether May-September is complete\",\"final rows are produced only after the final station, the first downstream failure, and the provisional downstream winner are all fixed under the same availability-page validation rule\"],\"control_dependency\":[\"station-level target-year Flow evidence must be validated by explicit May-September completeness checks on the availability page\",\"the final downstream reusable station and the first downstream failure must be determined under the same validation rule rather than from the initial station screen\",\"the provisional downstream winner is kept in the output only for reference after its FLOW_GAP status is confirmed on the availability page\"],\"freeze\":{\"historical_window\":\"Peace River mainstem stations between Dunvegan Bridge and Fifth Meridian, evaluated within the fixed historical 1966 and 1967 Flow-availability windows, with May-September complete-C coverage used as nearly stable reference information\"},\"answer_type\":\"multi-row ordered table\"}","rubric":"{\"inclusion_conditions\":[\"Screen only among Peace River mainstem stations between Dunvegan Bridge and Fifth Meridian.\",\"A genuinely reusable station must satisfy the availability-page rule that both 1966 and 1967 appear in the Flow table and that every month from May through September is complete.\",\"The final output must jointly report FINAL_STATION, FIRST_BREAK_AFTER_FINAL, and PROVISIONAL_STATION, while repeating provisional_count and final_count in every row.\"],\"exclusion_conditions\":[\"Exclude results that treat a station as usable simply because the target years have Flow records.\",\"Exclude results that do not verify completeness month by month from May through September.\",\"Exclude results that mark a station such as 07KA002 as FINAL_STATION when availability-page evidence shows FLOW_GAP.\"],\"normalization\":{\"field_separator\":\"|\",\"record_separator\":\"\\n\",\"header_row\":\"omit the header row; emit data rows only\",\"schema\":[\"rank\",\"station_number\",\"diagnosis_role\",\"target_window_status\",\"provisional_count\",\"final_count\"],\"rank\":\"emit the fixed order 1, 2, 3 for FINAL_STATION, FIRST_BREAK_AFTER_FINAL, and PROVISIONAL_STATION\",\"station_number\":\"use uppercase WSC station codes exactly as the Reference Index and availability pages show\",\"diagnosis_role\":[\"FINAL_STATION\",\"FIRST_BREAK_AFTER_FINAL\",\"PROVISIONAL_STATION\"],\"target_window_status\":[\"PASS\",\"LEVEL_ONLY\",\"FLOW_GAP\",\"NO_TARGET_RECORD\"],\"count_fields\":\"repeat the global provisional_count and final_count in every row as plain integers\",\"sorting_or_selection\":\"emit rows in the fixed diagnostic order FINAL_STATION, FIRST_BREAK_AFTER_FINAL, PROVISIONAL_STATION\"}}","all_involved_urls":"null"} |
| {"task_id":"waterquality_001","domain":"WATER_QUALITY_PORTAL","autonomy_type":"ordered table","oracle_output_cardinality":2,"instruction":"I am reviewing historical monitoring records for a reach of the middle Rio Grande and need to select two monitoring stations suitable for observing wastewater impacts on the main channel. Starting from RIO GRANDE AT EMBUDO, NM and proceeding downstream, retain only stations that are confirmed to lie on the Rio Grande mainstem. Then, find the first mainstem monitoring station downstream of the Town of Bernalillo wastewater treatment plant outfall that meets the following conditions; continuing further downstream, find the first mainstem monitoring station that has passed both the Town of Bernalillo WWTP outfall and the Rio Rancho No. 2 WWTP outfall and meets the same conditions. Both stations must satisfy: the station itself is not an outfall, facility, well, canal, or drainage ditch; the type on the station detail page is labeled as River/Stream; records for the four characteristic groups Microbiological, Nutrient, Organics, Other, and Physical fully cover 2015–2020; and the start years for both Microbiological and Physical records are earlier than 2010. Finally, output a single-line string in upstream-to-downstream order: <mainstem station ID>|<Microbiological start year>|<Physical start year>, with the two segments separated by a comma.","start_url":"https://www.waterqualitydata.us/","output_format":"Output a single-line string in upstream-to-downstream order: <mainstem station ID>|<Microbiological start year>|<Physical start year>, with the two segments separated by a comma.","oracle_answer":"SANDIAWQ_WQX-RG003|2004|2004,SANDIAWQ_WQX-RG008|2004|2004","metadata":"{\"State-Gated Retrieval\":[\"Select two target stations only from Rio Grande mainstem monitoring stations downstream of RIO GRANDE AT EMBUDO, NM, and both must have passed the specified outfall locations.\",\"The station detail page must explicitly show River/Stream, and records for the four characteristic groups Microbiological, Nutrient, Organics, Other, and Physical must fully cover 2015–2020, while the start years for both Microbiological and Physical must be earlier than 2010.\",\"Output the two final stations in upstream-to-downstream order.\"],\"dependency_type\":\"Data + Control\",\"intra_chain\":true,\"inter_chain\":true,\"data_dependency\":[\"provider/list views yield the Rio Grande mainstem candidates downstream of the Bernalillo and Rio Rancho No. 2 discharge points\",\"detail pages yield flow position, location type, characteristic-group period-of-record coverage for 2015-2020, and the Microbiological/Physical baseline years\",\"final rows are produced only after the corrected downstream sentry stations are fixed and ordered from upstream to downstream\"],\"control_dependency\":[\"provider-list stream tags are only provisional and must be confirmed on the detail page by both mainstem placement and location type\",\"target-window presence is insufficient until full 2015-2020 coverage is confirmed on the detail page for all four groups\",\"once RG002/RG007 fail the full detail-page checks, the workflow must continue downstream from the discharge-point starting position to RG003/RG008\"],\"freeze\":{\"historical_window\":\"Rio Grande mainstem stations downstream of the Town of Bernalillo and Rio Rancho No. 2 WWTP outfalls, validated against 2015-2020 coverage for Microbiological, Nutrient, Organics, Other, and Physical\"},\"answer_type\":\"multi-row ordered table\"}","rubric":"{\"inclusion_conditions\":[\"Select two target stations only from Rio Grande mainstem monitoring stations downstream of RIO GRANDE AT EMBUDO, NM, and both must have passed the specified outfall locations.\",\"The station detail page must explicitly show River/Stream, and records for the four characteristic groups Microbiological, Nutrient, Organics, Other, and Physical must fully cover 2015–2020, while the start years for both Microbiological and Physical must be earlier than 2010.\",\"Output the two final stations in upstream-to-downstream order.\"],\"exclusion_conditions\":[\"Exclude stations that have only a superficial (Stream) tag on the provider list or have records near the target window, but whose detail page does not satisfy the flow position, location type, or complete period-of-record conditions.\",\"Exclude locations that are outfalls, facilities, wells, canals, or drainage ditches.\",\"Exclude candidates such as RG002 and RG007 that are not the qualifying sentry stations downstream of the corresponding outfalls.\"],\"normalization\":{\"record_separator\":\",\",\"field_separator\":\"|\",\"schema_per_record\":[\"MonitoringLocationIdentifier\",\"Microbiological_start_year\",\"Physical_start_year\"],\"ordering\":\"emit exactly two records from upstream to downstream\",\"station_id\":\"use uppercase MonitoringLocationIdentifier\",\"year_fields\":\"emit four-digit years as plain integers\",\"selection\":{\"record_1\":\"the first qualifying mainstem station below the Bernalillo outfall\",\"record_2\":\"the first qualifying mainstem station below both the Bernalillo and Rio Rancho No. 2 outfalls\"},\"stop_condition\":[\"the Bernalillo branch stops only after its first qualifying mainstem station is fixed\",\"the Rio Rancho No. 2 branch stops only after its first qualifying mainstem station is fixed\",\"the two segments are emitted as one comma-joined line in upstream-to-downstream order\"]}}","all_involved_urls":"null"} |
| {"task_id":"waterquality_002","domain":"WATER_QUALITY_PORTAL","autonomy_type":"ordered table","oracle_output_cardinality":2,"instruction":"I am analyzing trends in the lower Rio Grande from 2001 to 2010. From monitoring stations located below RIO GRANDE BELOW ANZALDUAS and still within HUC 13090002, select one representative station for Hidalgo County and one for Cameron County. First, screen candidate stations for each county within the specified area. Retain only stations whose detail page lists the type as River/Stream and that have records covering 2001–2010 for all four characteristic groups: Microbiological, Nutrient, Organics, Other, and Physical. If multiple stations in the same county qualify, select the station with the earliest year in which all four groups are simultaneously present. Then record the MonitoringLocationIdentifier of the selected station, the earliest year all four groups are simultaneously present, and the start year for each of Microbiological, Nutrient, Organics, Other, and Physical. Finally, output a single-line string in the order Hidalgo then Cameron: <County>:<MonitoringLocationIdentifier>|<common_start_year>|<Microbiological_start_year>|<Nutrient_start_year>|<Organics_Other_start_year>|<Physical_start_year>, with the two segments separated by a comma.","start_url":"https://www.waterqualitydata.us/","output_format":"Output a single-line string in the order Hidalgo then Cameron: <County>:<MonitoringLocationIdentifier>|<common_start_year>|<Microbiological_start_year>|<Nutrient_start_year>|<Organics_Other_start_year>|<Physical_start_year>, with the two segments separated by a comma.","oracle_answer":"Hidalgo:TCEQMAIN-15808|1998|1998|1998|1998|1998,Cameron:TCEQMAIN-13177|1995|1981|1981|1995|1981","metadata":"{\"State-Gated Retrieval\":[\"Select one representative station for Hidalgo County and one for Cameron County from monitoring stations located below RIO GRANDE BELOW ANZALDUAS and still within HUC 13090002.\",\"Candidate detail pages must list the type as River/Stream, and records for all four characteristic groups (Microbiological, Nutrient, Organics, Other, Physical) must cover 2001–2010.\",\"If multiple stations in the same county qualify, select the station with the earliest year in which all four groups are simultaneously present.\"],\"dependency_type\":\"Data + Control\",\"intra_chain\":true,\"inter_chain\":true,\"data_dependency\":[\"the fixed downstream starting point and county split yield Hidalgo and Cameron candidate mainstem stations within HUC 13090002\",\"detail pages yield River/Stream validity, the four characteristic-group start years, and the earliest year when all four groups are simultaneously present for 2001-2010 coverage\",\"final rows are produced only after the county-specific searches are rerun from the fixed downstream starting point and then ranked by common-start year\"],\"control_dependency\":[\"the Hidalgo search must preserve the \\\"below TCEQMAIN-13664\\\" downstream starting point rather than returning to the anchor station itself\",\"a station is only provisional until all four characteristic groups are confirmed for the same MonitoringLocationIdentifier and tidal/intake conflicts are excluded on the detail page\",\"county winners are valid only after late-start and non-river candidates are removed and the common-start-year tie-break is recomputed\"],\"freeze\":{\"historical_window\":\"stations below RIO GRANDE BELOW ANZALDUAS within HUC 13090002, split by Hidalgo and Cameron County, with Microbiological, Nutrient, Organics, Other, and Physical coverage over 2001-2010\"},\"answer_type\":\"multi-row ordered table\"}","rubric":"{\"inclusion_conditions\":[\"Select one representative station for Hidalgo County and one for Cameron County from monitoring stations located below RIO GRANDE BELOW ANZALDUAS and still within HUC 13090002.\",\"Candidate detail pages must list the type as River/Stream, and records for all four characteristic groups (Microbiological, Nutrient, Organics, Other, Physical) must cover 2001–2010.\",\"If multiple stations in the same county qualify, select the station with the earliest year in which all four groups are simultaneously present.\"],\"exclusion_conditions\":[\"Exclude results where the Hidalgo search loses the downstream starting point and brings the starting station itself back into the candidate set.\",\"Exclude results that treat a single multi-select of the four characteristic groups as sufficient without verifying that all four belong to the same station ID.\",\"Exclude stations with tidal, intake, late-start, or other conflicts clearly indicated on the detail page.\"],\"normalization\":{\"record_separator\":\",\",\"field_separator\":\"|\",\"schema_per_record\":[\"County:MonitoringLocationIdentifier\",\"common_start_year\",\"Microbiological_start_year\",\"Nutrient_start_year\",\"Organics_Other_start_year\",\"Physical_start_year\"],\"ordering\":\"emit exactly two records, with Hidalgo first and Cameron second\",\"county_prefix\":\"use the county label in title case, followed by : and the uppercase MonitoringLocationIdentifier\",\"year_fields\":\"emit four-digit years as plain integers\",\"common_start_year\":\"the earliest year in which all four required groups are simultaneously present for the chosen county station\",\"stop_condition\":[\"each county branch is rerun until one River/Stream station survives all four group checks\",\"the county winner is fixed only after late-start and non-river candidates are excluded after checking the detail page\",\"the two county segments are emitted on one line in the fixed Hidalgo-then-Cameron order\"]}}","all_involved_urls":"null"} |
| {"task_id":"waterquality_003","domain":"WATER_QUALITY_PORTAL","autonomy_type":"ordered table","oracle_output_cardinality":3,"instruction":"I am compiling 2012–2018 water quality reference data for the Brazos River mainstem downstream of Brazos Rv at Waco, TX. I want to select one TCEQ monitoring station suitable as a long-term reference point in each of the three HUCs (12060202, 12070101, 12070104) that the mainstem passes through. Process as follows: first, within the Brazos River mainstem reach downstream of Brazos Rv at Waco, TX, screen candidate stations in HUCs 12060202, 12070101, and 12070104 respectively; retain only monitoring locations whose detail page labels them as River/Stream and whose station name still corresponds to the Brazos River mainstem rather than another named waterbody. Next, check the record categories for these stations, retaining only those that cover 2012–2018 for all five categories: Inorganics, Major, Non-metals; Microbiological; Nutrient; Organics, Other; and Physical. Specifically, the station start year must be no later than 2012 and the end year no earlier than 2018. Finally, list the 3 results in order from upstream to downstream, each row formatted as: <HUC>|<MonitoringLocationIdentifier>|<earliest year when all five categories are jointly available>|<Inorganics, Major, Non-metals start year>|<Microbiological start year>|<Nutrient start year>|<Organics, Other start year>|<Physical start year>.","start_url":"https://www.waterqualitydata.us/","output_format":"List the 3 results in order from upstream to downstream, each row formatted as: <HUC>|<MonitoringLocationIdentifier>|<earliest year when all five categories are jointly available>|<Inorganics, Major, Non-metals start year>|<Microbiological start year>|<Nutrient start year>|<Organics, Other start year>|<Physical start year>.","oracle_answer":"12060202|TCEQMAIN-12037|2000|1991|2000|1991|1999|1991\n12070101|TCEQMAIN-12030|2006|1972|1972|1972|2006|1972\n12070104|TCEQMAIN-11848|1975|1974|1975|1974|1975|1974","metadata":"{\"State-Gated Retrieval\":[\"The search must stay anchored downstream of Waco and look for Brazos mainstem monitoring stations within the three HUCs specified in the task.\",\"Each candidate station must be checked separately across the five characteristic groups and then matched back to the same station ID, after which the detail page must verify full 2012–2018 coverage and mainstem semantics.\",\"For each HUC, retain only the qualifying station with the earliest common start year, and output the start-year fields required by the task.\"],\"dependency_type\":\"Data + Control\",\"intra_chain\":true,\"inter_chain\":true,\"data_dependency\":[\"the fixed downstream starting point below Waco and the three target HUC searches yield the Brazos mainstem candidate stations\",\"detail pages yield true downstream position, mainstem-vs-tributary semantics, and five-group period-of-record coverage across 2012-2018\",\"final rows are produced only after the three HUC-specific searches are carried out, matched by station ID, and ranked by the earliest year when all five groups are jointly available\"],\"control_dependency\":[\"the workflow must stay anchored downstream of Waco rather than reintroducing upstream stations such as TCEQMAIN-12044\",\"name-based waterbody matches must be validated on the detail page by excluding tributaries such as LITTLE BRAZOS RIVER and confirming full-window coverage\",\"station ranking must be determined after false positives from partial-window records are removed\"],\"freeze\":{\"historical_window\":\"Brazos River mainstem stations downstream of the Waco anchor, split across HUC 12060202, 12070101, and 12070104, with five characteristic groups covering 2012-2018\"},\"answer_type\":\"multi-row ordered table\"}","rubric":"{\"inclusion_conditions\":[\"The search must stay anchored downstream of Waco and look for Brazos mainstem monitoring stations within the three HUCs specified in the task.\",\"Each candidate station must be checked separately across the five characteristic groups and then matched back to the same station ID, after which the detail page must verify full 2012–2018 coverage and mainstem semantics.\",\"For each HUC, retain only the qualifying station with the earliest common start year, and output the start-year fields required by the task.\"],\"exclusion_conditions\":[\"Exclude results that lose the downstream starting point below Waco and bring upstream stations back into the candidate set.\",\"Exclude results that mistake tributaries such as LITTLE BRAZOS RIVER for the Brazos mainstem.\",\"Exclude stations that have data near the query window but do not cover the entire 2012–2018 window.\"],\"normalization\":{\"field_separator\":\"|\",\"record_separator\":\"\\n\",\"header_row\":\"omit the header row; emit data rows only\",\"schema\":[\"HUC\",\"MonitoringLocationIdentifier\",\"common_start_year\",\"Inorganics_start\",\"Microbiological_start\",\"Nutrient_start\",\"Organics_Other_start\",\"Physical_start\"],\"HUC\":\"preserve the HUC exactly as emitted in the source result\",\"MonitoringLocationIdentifier\":\"preserve the monitoring-location identifier exactly, for example `TCEQMAIN-12037`\",\"common_start_year\":\"emit the maximum of the five per-group start years\",\"selection_rule\":\"only keep mainstem river/stream sites whose five required groups all cover the target 2012-2018 window\",\"sorting_or_selection\":\"upstream to downstream along the mainstem\"}}","all_involved_urls":"null"} |
| {"task_id":"waterquality_004","domain":"WATER_QUALITY_PORTAL","autonomy_type":"ordered table","oracle_output_cardinality":3,"instruction":"I am reviewing toxic-pollutant-related water quality monitoring data for 1993–2001 along the Trinity River mainstem, bounded upstream by the reach near FM 85 and downstream by the reach near Interstate 10. I need to select one TCEQ long-term reference station for each of the upper, middle, and lower segments.\n\nStarting from the Trinity River mainstem between the reach near FM 85 and the reach near Interstate 10, screen candidate monitoring locations in the upper, middle, and lower segments. Retain only stations whose record is tagged as River/Stream and whose name still refers to the Trinity River mainstem, not to other named water bodies or distributaries. Then check the record categories for these stations; keep only those that have all six record categories covering 1993–2001 (i.e., the start year is no later than 1993 and the end year is no earlier than 2001). Output the results from upstream to downstream.\n\nFinally, list three rows from upstream to downstream along the river, each in the format: <HUC>|<MonitoringLocationIdentifier>|<earliest year with all six categories>|<Inorganics, Major, Non-metals start year>|<Microbiological start year>|<Nutrient start year>|<Organics, Other start year>|<Organics, PCBs start year>|<Physical start year>.","start_url":"https://www.waterqualitydata.us/","output_format":"List three rows from upstream to downstream along the river, each in the format: <HUC>|<MonitoringLocationIdentifier>|<earliest year with all six categories>|<Inorganics, Major, Non-metals start year>|<Microbiological start year>|<Nutrient start year>|<Organics, Other start year>|<Organics, PCBs start year>|<Physical start year>.","oracle_answer":"12030105|TCEQMAIN-10924|1983|1982|1983|1982|1983|1983|1982\n12030202|TCEQMAIN-10897|1993|1969|1972|1972|1993|1974|1969\n12030203|TCEQMAIN-10892|1976|1969|1971|1972|1974|1976|1969","metadata":"{\"State-Gated Retrieval\":[\"Screen only among candidate stations where Trinity River mainstem semantics hold, and require all six characteristic groups to cover 1993–2001.\",\"The six groups must be searched separately and then matched back to the same station ID, after which the detail page must verify full-window coverage and legacy toxics group completeness.\",\"For each target segment, retain only the station that meets all group requirements and has the best common start year.\"],\"dependency_type\":\"Data + Control\",\"intra_chain\":true,\"inter_chain\":true,\"data_dependency\":[\"Trinity River mainstem candidate stations are generated from the target HUC searches and the six characteristic-group searches\",\"detail pages yield River/Stream validity, 1993–2001 full-window coverage, and the group-specific start years needed for ranking\",\"final rows are produced only after all six groups have been matched by station ID and detail-page validation removes partial-group and non-mainstem false positives\"],\"control_dependency\":[\"each of the six characteristic groups must be evaluated separately and confirmed against the same station ID before a multi-select result can be retained\",\"station names and River/Stream tags are insufficient until the detail page confirms Trinity mainstem semantics and 1993–2001 full-window coverage\",\"candidate rejection and final ranking must be determined after applying the PCBs / Organics, Other gap checks\"],\"freeze\":{\"historical_window\":\"Trinity River mainstem stations validated against six characteristic groups over 1993–2001, with ranking based on the earliest joint availability and group start years\"},\"answer_type\":\"multi-row ordered table\"}","rubric":"{\"inclusion_conditions\":[\"Screen only among candidate stations where Trinity River mainstem semantics hold, and require all six characteristic groups to cover 1993–2001.\",\"The six groups must be searched separately and then matched back to the same station ID, after which the detail page must verify full-window coverage and legacy toxics group completeness.\",\"For each target segment, retain only the station that meets all group requirements and has the best common start year.\"],\"exclusion_conditions\":[\"Exclude results that mistakenly treat a single multi-select of characteristic groups as sufficient to satisfy the requirement.\",\"Exclude results that only check station names and River/Stream tags without verifying 1993–2001 full-window coverage.\",\"Exclude stations that ignore missing PCBs or Organics, Other groups.\"],\"normalization\":{\"field_separator\":\"|\",\"record_separator\":\"\\n\",\"header_row\":\"omit the header row; emit data rows only\",\"schema\":[\"HUC\",\"MonitoringLocationIdentifier\",\"common_start_year\",\"Inorganics_start\",\"Microbiological_start\",\"Nutrient_start\",\"Organics_Other_start\",\"Organics_PCBs_start\",\"Physical_start\"],\"HUC\":\"preserve the HUC exactly as emitted in the source result\",\"MonitoringLocationIdentifier\":\"preserve the monitoring-location identifier exactly, for example `TCEQMAIN-10924`\",\"common_start_year\":\"emit the maximum of the six per-group start years\",\"selection_rule\":\"only keep mainstem river/stream sites whose six required groups all cover the target window\",\"sorting_or_selection\":\"upstream to downstream along the mainstem\"}}","all_involved_urls":"null"} |
| {"task_id":"waterquality_005","domain":"WATER_QUALITY_PORTAL","autonomy_type":"ordered table","oracle_output_cardinality":9,"instruction":"I am organizing integrated water-quality records for the lower Neches River during 1978-1984. Starting downstream from NECHES RIVER AT CR 4915, I want to identify the nearest TCEQ long-term reference station that still represents the inland Neches River mainstem, and then tabulate the start year of nine Characteristic Groups at that station. Search downstream along the mainstem and retain only monitoring stations whose detail page labels the site type as River/Stream. Exclude stations located in lake or reservoir waters or in the Sabine-Neches outlet boundary reach. Among stations that satisfy these location conditions, identify the nearest qualifying station for which all nine record categories cover 1978-1984: Biological, Algae, Phytoplankton; Inorganics, Major, Non-metals; Inorganics, Minor, Metals; Microbiological; Nutrient; Organics, Other; Organics, PCBs; Organics, Pesticide; and Physical. Finally, output a table for that final station, one row per Characteristic Group, with columns rank|MonitoringLocationIdentifier|HUC|County|characteristic_group|group_start_year|common_start_year. The row order is fixed as Biological, Algae, Phytoplankton; Inorganics, Major, Non-metals; Inorganics, Minor, Metals; Microbiological; Nutrient; Organics, Other; Organics, PCBs; Organics, Pesticide; Physical.","start_url":"https://www.waterqualitydata.us/provider/STORET/TCEQMAIN/TCEQMAIN-10598/","output_format":"Output a 9-row table with columns rank|MonitoringLocationIdentifier|HUC|County|characteristic_group|group_start_year|common_start_year.","oracle_answer":"1|TCEQMAIN-10580|12020003|Hardin County|Biological, Algae, Phytoplankton|1972|1977\n2|TCEQMAIN-10580|12020003|Hardin County|Inorganics, Major, Non-metals|1971|1977\n3|TCEQMAIN-10580|12020003|Hardin County|Inorganics, Minor, Metals|1974|1977\n4|TCEQMAIN-10580|12020003|Hardin County|Microbiological|1972|1977\n5|TCEQMAIN-10580|12020003|Hardin County|Nutrient|1972|1977\n6|TCEQMAIN-10580|12020003|Hardin County|Organics, Other|1974|1977\n7|TCEQMAIN-10580|12020003|Hardin County|Organics, PCBs|1974|1977\n8|TCEQMAIN-10580|12020003|Hardin County|Organics, Pesticide|1977|1977\n9|TCEQMAIN-10580|12020003|Hardin County|Physical|1971|1977","metadata":"{\"State-Gated Retrieval\":[\"Search only among TCEQ monitoring stations downstream of NECHES RIVER AT CR 4915 that still represent the inland Neches River mainstem.\",\"A candidate station must cover 1978-1984 in all nine Characteristic Groups and, on the detail page, must still qualify as River/Stream, outside lake or reservoir waters, and outside the Sabine-Neches outlet boundary reach.\",\"The final output must be centered on the single winning station and list group_start_year plus common_start_year row by row in the fixed nine-group order.\"],\"dependency_type\":\"Data + Control\",\"intra_chain\":true,\"inter_chain\":true,\"data_dependency\":[\"the fixed downstream starting point at NECHES RIVER AT CR 4915 yields candidate inland mainstem stations across successive downstream HUCs\",\"detail pages yield River/Stream validity, inland-vs-boundary semantics, and nine-group period-of-record coverage across 1978-1984\",\"final rows are produced only after one winning station survives the inland-mainstem checks and the requirement that all nine groups match the same station, after which its nine group start years are emitted in the required order\"],\"control_dependency\":[\"the nearest downstream mainstem station remains provisional until all nine characteristic groups have been checked separately and matched back to the same station ID\",\"the downstream search must continue to the next mainstem HUC when the first candidate fails the nine-group coverage requirement\",\"location semantics remain valid only after detail-page exclusion of lake/reservoir and Sabine-Neches boundary stations\"],\"freeze\":{\"historical_window\":\"Neches River inland mainstem stations downstream of NECHES RIVER AT CR 4915, validated within the fixed 1978-1984 window against nine characteristic groups as largely stable reference information\"},\"answer_type\":\"multi-row ordered table\"}","rubric":"{\"inclusion_conditions\":[\"Search only among TCEQ monitoring stations downstream of NECHES RIVER AT CR 4915 that still represent the inland Neches River mainstem.\",\"A candidate station must cover 1978-1984 in all nine Characteristic Groups and, on the detail page, must still qualify as River/Stream, outside lake or reservoir waters, and outside the Sabine-Neches outlet boundary reach.\",\"The final output must be centered on the single winning station and list group_start_year plus common_start_year row by row in the fixed nine-group order.\"],\"exclusion_conditions\":[\"Exclude results that retain TCEQMAIN-10591 merely because it appears to be the nearest downstream mainstem station.\",\"Exclude results that do not check the nine Characteristic Groups separately and verify that they all belong to the same station ID.\",\"Exclude stations located in lake or reservoir waters or in the Sabine-Neches outlet boundary reach.\"],\"normalization\":{\"field_separator\":\"|\",\"record_separator\":\"\\n\",\"header_row\":\"omit the header row; emit data rows only\",\"schema\":[\"rank\",\"MonitoringLocationIdentifier\",\"HUC\",\"County\",\"characteristic_group\",\"group_start_year\",\"common_start_year\"],\"rank\":\"emit 1 through 9 in the fixed characteristic-group order required by the task\",\"MonitoringLocationIdentifier\":\"use uppercase MonitoringLocationIdentifier\",\"HUC\":\"emit the HUC as an 8-digit string with no separators\",\"County\":\"use the county name exactly as shown on the detail page\",\"characteristic_group\":\"preserve the group label exactly as listed in the task\",\"year_fields\":\"emit four-digit years as plain integers\",\"common_start_year\":\"repeat the earliest year in which all nine required groups are simultaneously present for the chosen station\",\"sorting_or_selection\":\"emit rows in the fixed nine-group order stated in the task\"}}","all_involved_urls":"null"} |
| {"task_id":"wonder_001","domain":"CDC_WONDER","autonomy_type":"ordered table","oracle_output_cardinality":8,"instruction":"I am preparing a public health brief on changes in U.S. home births in 2020. I need to identify months where the home-birth rate among all mothers increased by at least 25% compared to the same month in 2019, and also examine the corresponding changes for non-Hispanic Black mothers and Hispanic mothers in those months. First, compare the home-birth rates for all mothers by calendar month between 2019 and 2020, retaining only months in 2020 where the rate increased by at least 25% relative to the same month in 2019. Then, for those months, add the 2020 home-birth rates for non-Hispanic Black mothers and Hispanic mothers, along with their changes relative to the same month in 2019. Finally, output a multi-row pipe-delimited table in calendar month order with the fixed header: month|all_2019|all_2020|all_rel|black_2020|black_rel|hispanic_2020|hispanic_rel. All numeric values should be plain numbers without percent signs.","start_url":"https://wonder.cdc.gov/natality-expanded-current.html","output_format":"Output a multi-row pipe-delimited table in calendar month order with the fixed header: month|all_2019|all_2020|all_rel|black_2020|black_rel|hispanic_2020|hispanic_rel. All numeric values should be plain numbers without percent signs.","oracle_answer":"month|all_2019|all_2020|all_rel|black_2020|black_rel|hispanic_2020|hispanic_rel\nMay|1.03|1.49|45|0.74|54|0.55|57\nJune|1.04|1.31|26|0.66|20|0.48|30\nJuly|1.03|1.30|26|0.68|42|0.48|30\nAugust|1.02|1.29|26|0.69|50|0.48|26\nSeptember|1.01|1.31|30|0.73|43|0.54|54\nOctober|1.00|1.28|28|0.75|50|0.47|27\nNovember|0.99|1.30|31|0.73|43|0.55|67\nDecember|0.92|1.28|39|0.80|63|0.53|61","metadata":"{\"State-Gated Retrieval\":[\"First, compare the home-birth rates for all mothers by calendar month between 2019 and 2020, retaining only months in 2020 where the rate increased by at least 25% relative to the same month in 2019.\",\"The home-birth numerator must use the full home category, not Home intended alone.\",\"For those months, add the 2020 home-birth rates for non-Hispanic Black and Hispanic mothers, along with their changes relative to the same month in 2019.\"],\"dependency_type\":\"Data + Control\",\"intra_chain\":true,\"inter_chain\":true,\"data_dependency\":[\"Monthly natality queries yield the 2019 vs 2020 home-birth rates for all mothers and the candidate month set.\",\"Follow-up race/ethnicity queries yield the 2020 home-birth rates for non-Hispanic Black and Hispanic mothers, plus their same-month changes versus 2019.\",\"Final rows are produced only after qualifying months are determined from rate growth, and the resulting month set is reused for both subgroup backfill chains.\"],\"control_dependency\":[\"Month selection must be based on relative rate increase rather than counts alone.\",\"The home-birth numerator must use Residence or an equivalent full-home measure rather than Home intended alone before downstream subgroup queries run.\",\"Subgroup columns must be generated from the resulting month set rather than appended to earlier rows.\"],\"freeze\":{\"historical_window\":\"2019 and 2020 U.S. natality monthly home-birth rates for all mothers, non-Hispanic Black mothers, and Hispanic mothers.\"},\"answer_type\":\"multi-row ordered table\"}","rubric":"{\"inclusion_conditions\":[\"First, compare the home-birth rates for all mothers by calendar month between 2019 and 2020, retaining only months in 2020 where the rate increased by at least 25% relative to the same month in 2019.\",\"The home-birth numerator must use the full home category, not Home intended alone.\",\"For those months, add the 2020 home-birth rates for non-Hispanic Black and Hispanic mothers, along with their changes relative to the same month in 2019.\"],\"exclusion_conditions\":[\"Exclude month-row sets that misinterpret \\\"rate increase of at least 25%\\\" as \\\"count increase of at least 25%.\\\"\",\"Exclude results that use only the Home intended category instead of the full home numerator.\",\"Exclude results that add subgroup columns to an incorrect month set.\"],\"normalization\":{\"field_separator\":\"|\",\"record_separator\":\"\\n\",\"header_row\":\"Required literal first row: month|all_2019|all_2020|all_rel|black_2020|black_rel|hispanic_2020|hispanic_rel\",\"schema\":[\"month\",\"all_2019\",\"all_2020\",\"all_rel\",\"black_2020\",\"black_rel\",\"hispanic_2020\",\"hispanic_rel\"],\"month\":\"Use the English month name and keep calendar order.\",\"numeric_fields\":\"All numeric outputs are plain numbers without a percent sign.\",\"relative_change_fields\":\"all_rel, black_rel, and hispanic_rel are emitted as integer percentages without the `%` character.\"}}","all_involved_urls":"null"} |
| {"task_id":"wonder_002","domain":"CDC_WONDER","autonomy_type":"ordered table","oracle_output_cardinality":3,"instruction":"I am preparing a one-page public health brief on 2021 U.S. infant mortality disparities. I want to identify groups that are separately listed in the report by mother's race and Hispanic origin, have an infant mortality rate above the U.S. overall rate, and have a clearly identifiable leading cause of death. First, from the separately listed groups in the report, select those with an infant mortality rate above the U.S. overall rate and a clearly identifiable leading cause of death. Then, for each qualifying group, record the full-year infant mortality rate, the neonatal rate (under 28 days), and the postneonatal rate (28 days and older), along with the group's leading cause of death, its cause-specific rate, and the rank of that cause among all infant causes. Finally, output a multi-row pipe-delimited table sorted by full-year infant mortality rate from high to low: group|infant_rate|neonatal_rate|postneonatal_rate|top_cause|cause_rate|us_rank.","start_url":"https://wonder.cdc.gov/lbd-current-expanded.html","output_format":"Finally, output a multi-row pipe-delimited table sorted by full-year infant mortality rate from high to low: group|infant_rate|neonatal_rate|postneonatal_rate|top_cause|cause_rate|us_rank.","oracle_answer":"group|infant_rate|neonatal_rate|postneonatal_rate|top_cause|cause_rate|us_rank\nnon-Hispanic Black|10.55|6.35|4.19|Short gestation and low birth weight|196.4|2\nnon-Hispanic American Indian or Alaska Native|7.46|3.83|3.67|Congenital malformations|134.0|1\nPuerto Rican|6.05|3.93|2.13|Short gestation and low birth weight|108.9|2","metadata":"{\"State-Gated Retrieval\":[\"Retain only groups that are separately listed in the 2021 U.S. infant mortality statistics, have a full-year infant mortality rate above the U.S. overall rate, and have a clearly identifiable leading cause of death.\",\"For each qualifying group, provide the infant rate, neonatal rate, postneonatal rate, top cause, cause rate, and the rank of that cause among all infant causes.\",\"The leading cause must be determined by a group-level leading-cause query and verified for availability using notes/reliability markers.\"],\"dependency_type\":\"Data + Control\",\"intra_chain\":true,\"inter_chain\":true,\"data_dependency\":[\"infant-mortality queries yield group-level overall, neonatal, and postneonatal rates for reportable mother race/Hispanic-origin groups\",\"leading-cause queries and reliability notes yield each eligible group's top cause, cause-specific rate, and U.S. rank\final rows are produced only after the eligible-group set is rebuilt using both overall-rate and cause-specific-identifiability criteria\control_dependency\groups above the U.S. overall rate cannot be accepted until cause-specific reliability checks are run, which removes NHOPI\leading cause must be queried at the group level rather than inherited from an all-infants cause table\the cause-rate multiplier must be defined as per 100,000 live births before downstream ranking/output\freeze\historical_window\2021 U.S. infant mortality by mother race and Hispanic origin, including the group-level leading-cause tables and reliability notes\answer_type\multi-row ordered table\,:inclusion_conditions\Retain only groups that are separately listed in the 2021 U.S. infant mortality statistics, have a full-year infant mortality rate above the U.S. overall rate, and have a clearly identifiable leading cause of death.\For each qualifying group, provide the infant rate, neonatal rate, postneonatal rate, top cause, cause rate, and the rank of that cause among all infant causes.\The leading cause must be determined by a group-level leading-cause query and verified for availability using notes/reliability markers.\exclusion_conditions\Exclude groups that, despite having a full-year mortality rate above the U.S. overall rate, cannot provide a clearly identifiable leading cause due to cause-specific reliability issues.\Exclude results that only inherit the top cause from the all-infants cause table without performing a group-level leading-cause query.\Exclude results where the cause_rate uses an incorrect rate multiplier and is not read at the per-100,000-live-births denominator.\normalization\field_separator\|\record_separator\\\n\header_row\required literal first row `group|infant_rate|neonatal_rate|postneonatal_rate|top_cause|cause_rate|us_rank`\",\"schema\":[\"group\",\"infant_rate\",\"neonatal_rate\",\"postneonatal_rate\",\"top_cause\",\"cause_rate\",\"us_rank\"],\"sorting_or_selection\":\"infant_rate descending\",\"rate_fields\":\"infant_rate, neonatal_rate, and postneonatal_rate are per 1,000 live births; cause_rate is per 100,000 live births, but all four are emitted as plain numbers without unit text\",\"top_cause\":\"emit the short cause label used in the oracle, not an expanded prose description\",\"us_rank\":\"emit the integer rank of that same cause in the overall infant leading-cause list\"}}","all_involved_urls":"null"} |
|
|
