Dataset Viewer
Auto-converted to Parquet Duplicate
question
stringlengths
687
756
answer
stringclasses
1 value
task_config
stringclasses
1 value
task_source
stringclasses
1 value
prompt
listlengths
1
1
model_responses
listlengths
0
0
model_responses__eval_is_correct
listlengths
0
0
all_other_columns
stringlengths
162
231
original_split
stringclasses
1 value
acronym
stringlengths
4
4
formed_acronym
stringlengths
4
4
word_count
int64
4
8
words
listlengths
4
8
model_responses__best_of_n_atags
listlengths
4
4
model_responses__best_of_n_atags__finish_reason_length_flags
listlengths
4
4
model_responses__best_of_n_atags__length_partial_responses
listlengths
4
4
prompt__best_of_n_atags__metadata
dict
model_responses__best_of_n_atags__metadata
dict
model_responses__best_of_n_atags__eval_is_correct
listlengths
4
4
model_responses__best_of_n_atags__eval_extracted_answers
listlengths
4
4
model_responses__best_of_n_atags__eval_extraction_metadata
stringlengths
350
594
model_responses__best_of_n_atags__eval_evaluation_metadata
stringlengths
384
680
model_responses__best_of_n_atags__internal_answers__eval_is_correct
listlengths
4
4
model_responses__best_of_n_atags__internal_answers__eval_extracted_answers
listlengths
4
4
model_responses__best_of_n_atags__internal_answers__eval_extraction_metadata
stringlengths
932
17k
model_responses__best_of_n_atags__internal_answers__eval_evaluation_metadata
stringlengths
492
9.43k
model_responses__best_of_n_atags__metrics
dict
eval_date
stringdate
2025-11-27 00:00:00
2025-11-27 00:00:00
split
stringclasses
1 value
revision_name
stringclasses
1 value
model_path
stringclasses
1 value
checkpoint_step
int64
0
0
stage_name
stringclasses
1 value
stage_number
int64
1
1
timestamp
stringdate
2025-11-27 21:20:34
2025-11-27 21:20:34
eval_repo_id
stringclasses
1 value
Answer the following problem. Explain your reasoning step by step. When you are finished, give your answer in this format: <answer> (your answer) </answer>. ### Problem You are given a list of words. Construct an acronym that satisfies the following conditions: 1) is at least 4 characters 2) uses the first letter...
acronym_4o
acronym
[ { "content": "Answer the following problem. Explain your reasoning step by step. When you are finished, give your answer in this format: <answer> (your answer) </answer>.\n\n ### Problem\n\n\nYou are given a list of words. Construct an acronym that satisfies the following conditions:\n\n1) is at least 4 charac...
[]
[]
{"acronym": "they", "words": ["treating", "hurricane", "exhibits", "yours"], "word_count": 4, "formed_acronym": "they", "task_config": "acronym_4o", "task_source": "acronym"}
test
they
they
4
[ "treating", "hurricane", "exhibits", "yours" ]
[ "<think>\n<sample>\nTo construct an acronym that meets the given conditions, we need to follow these steps:\n\n1. **Identify the required length**: The acronym must be at least 4 characters long.\n2. **Select starting letters**: We need to use the first letter of at least 4 words from the list.\n3. **Form a valid w...
[ false, false, false, false ]
[ "", "", "", "" ]
{ "api_url": "http://10.32.36.236:9080/v1", "backend": "litellm", "chat_template_applied": true, "generation_params": { "chat_template_applied": true, "max_tokens": 16384, "n": 1, "repetition_penalty": 1.1, "temperature": 0.7, "top_k": 20, "top_p": 0.8 }, "model_name": "hosted_vl...
{ "backend": "litellm", "model_name": "hosted_vllm/TAUR-dev/M-1119_newmodels__qwen7b_sft1k_ours-rl@checkpoint-step-150", "n_responses": 4 }
[ true, true, true, true ]
[ "they", "they", "they", "they" ]
[{"extraction_type": "final", "total_spans_found": 50, "non_overlapping_spans": 50, "span_start": 64381, "span_end": 64387, "original_span": "they"}, {"extraction_type": "final", "total_spans_found": 6, "non_overlapping_spans": 6, "span_start": 10199, "span_end": 10205, "original_span": "they"}, {"extraction_type": "fi...
[{"is_correct": true, "error": "", "answer_block": "they", "final_answer": "they", "length": 4}, {"is_correct": true, "error": "", "answer_block": "they", "final_answer": "they", "length": 4}, {"is_correct": true, "error": "", "answer_block": "they", "final_answer": "they", "length": 4}, {"is_correct": true, "error": "...
[ [ false, true, true, false, true, true, true, true, true, true, true, true, true, true, true, true, true, true, true, true, true, true, true, true, true, true, true, true, true, true, true, ...
[ [ "thye", "they", "they", "there", "they", "they", "they", "they", "they", "they", "they", "they", "they", "they", "they", "they", "they", "they", "they", "they", "they", "they", "they", "they", "they", "they", ...
[[{"pattern_used": "answer tag", "confidence": 1.0, "position": 1504, "span_text": "thye", "extraction_type": "internal", "span_start": 1504, "span_end": 1510, "original_span": "thye"}, {"pattern_used": "answer tag", "confidence": 1.0, "position": 4252, "span_text": "they", "extraction_type": "internal", "span_start": ...
[[{"is_correct": false, "error": "Constraint violations: not_valid_english_word, violates_character_position_constraint", "answer_block": "thye", "final_answer": "thye", "length": 4}, {"is_correct": true, "error": "", "answer_block": "they", "final_answer": "they", "length": 4}, {"is_correct": true, "error": "", "answe...
{ "flips_by": [ 1, 0, 1, 0 ], "flips_total": 2, "num_correct": 4, "pass_at_n": 1, "percent_correct": 1, "total_responses": 4 }
2025-11-27
test
main
TAUR-dev/M-1119_newmodels__qwen7b_sft1k_ours-rl@checkpoint-step-150
0
eval_rl
1
2025-11-27T21:20:34.975869
TAUR-dev/D-EVAL__standard_eval_v3__FinEval_16k_fulleval_3arg_OT_ours_1k-at150-RL-acronym_4o-eval_rl
"Answer the following problem. Explain your reasoning step by step. When you are finished, give your(...TRUNCATED)
acronym_4o
acronym
[{"content":"Answer the following problem. Explain your reasoning step by step. When you are finishe(...TRUNCATED)
[]
[]
"{\"acronym\": \"when\", \"words\": [\"whether\", \"hooked\", \"enables\", \"next\"], \"word_count\"(...TRUNCATED)
test
when
when
4
[ "whether", "hooked", "enables", "next" ]
["<think>\n<sample>\nTo construct an acronym that meets all the given conditions, we need to follow (...TRUNCATED)
[ false, false, false, false ]
[ "", "", "", "" ]
{"api_url":"http://10.32.36.236:9080/v1","backend":"litellm","chat_template_applied":true,"generatio(...TRUNCATED)
{"backend":"litellm","model_name":"hosted_vllm/TAUR-dev/M-1119_newmodels__qwen7b_sft1k_ours-rl@check(...TRUNCATED)
[ true, true, true, true ]
[ "when", "when", "when", "when" ]
"[{\"extraction_type\": \"final\", \"total_spans_found\": 5, \"non_overlapping_spans\": 5, \"span_st(...TRUNCATED)
"[{\"is_correct\": true, \"error\": \"\", \"answer_block\": \"when\", \"final_answer\": \"when\", \"(...TRUNCATED)
[[true,true,true,true,true],[false,false,true,true,true,true],[false,false,true,false,true,true,true(...TRUNCATED)
[["when","when","when","when","when"],["won","wrenches","when","when","when","when"],["hewn","wet","(...TRUNCATED)
"[[{\"pattern_used\": \"answer tag\", \"confidence\": 1.0, \"position\": 1272, \"span_text\": \"when(...TRUNCATED)
"[[{\"is_correct\": true, \"error\": \"\", \"answer_block\": \"when\", \"final_answer\": \"when\", \(...TRUNCATED)
{"flips_by":[0,0,1,1],"flips_total":2,"num_correct":4,"pass_at_n":1,"percent_correct":1.0,"total_res(...TRUNCATED)
2025-11-27
test
main
TAUR-dev/M-1119_newmodels__qwen7b_sft1k_ours-rl@checkpoint-step-150
0
eval_rl
1
2025-11-27T21:20:34.975882
"TAUR-dev/D-EVAL__standard_eval_v3__FinEval_16k_fulleval_3arg_OT_ours_1k-at150-RL-acronym_4o-eval_rl(...TRUNCATED)
"Answer the following problem. Explain your reasoning step by step. When you are finished, give your(...TRUNCATED)
acronym_4o
acronym
[{"content":"Answer the following problem. Explain your reasoning step by step. When you are finishe(...TRUNCATED)
[]
[]
"{\"acronym\": \"some\", \"words\": [\"sufficient\", \"overwhelming\", \"mentally\", \"elizabeth\"],(...TRUNCATED)
test
some
some
4
[ "sufficient", "overwhelming", "mentally", "elizabeth" ]
["<think>\n<sample>\nTo construct an acronym that meets all the given conditions, we need to follow (...TRUNCATED)
[ false, false, false, false ]
[ "", "", "", "" ]
{"api_url":"http://10.32.36.236:9080/v1","backend":"litellm","chat_template_applied":true,"generatio(...TRUNCATED)
{"backend":"litellm","model_name":"hosted_vllm/TAUR-dev/M-1119_newmodels__qwen7b_sft1k_ours-rl@check(...TRUNCATED)
[ true, true, true, true ]
[ "some", "some", "some", "some" ]
"[{\"extraction_type\": \"final\", \"total_spans_found\": 6, \"non_overlapping_spans\": 6, \"span_st(...TRUNCATED)
"[{\"is_correct\": true, \"error\": \"\", \"answer_block\": \"some\", \"final_answer\": \"some\", \"(...TRUNCATED)
[[false,false,true,true,true,true],[false,false,false,true,true,true,true,true,true],[false,false,fa(...TRUNCATED)
[["sold","mental","some","some","some","some"],["sold","sold","sole","some","some","some","some","so(...TRUNCATED)
"[[{\"pattern_used\": \"answer tag\", \"confidence\": 1.0, \"position\": 1536, \"span_text\": \"sold(...TRUNCATED)
"[[{\"is_correct\": false, \"error\": \"Constraint violations: violates_character_position_constrain(...TRUNCATED)
{"flips_by":[0,0,1,0],"flips_total":1,"num_correct":4,"pass_at_n":1,"percent_correct":1.0,"total_res(...TRUNCATED)
2025-11-27
test
main
TAUR-dev/M-1119_newmodels__qwen7b_sft1k_ours-rl@checkpoint-step-150
0
eval_rl
1
2025-11-27T21:20:34.975884
"TAUR-dev/D-EVAL__standard_eval_v3__FinEval_16k_fulleval_3arg_OT_ours_1k-at150-RL-acronym_4o-eval_rl(...TRUNCATED)
"Answer the following problem. Explain your reasoning step by step. When you are finished, give your(...TRUNCATED)
acronym_4o
acronym
[{"content":"Answer the following problem. Explain your reasoning step by step. When you are finishe(...TRUNCATED)
[]
[]
"{\"acronym\": \"back\", \"words\": [\"believed\", \"allison\", \"clips\", \"kane\"], \"word_count\"(...TRUNCATED)
test
back
back
4
[ "believed", "allison", "clips", "kane" ]
["<think>\n<sample>\nTo construct an acronym that meets the given conditions, we need to follow thes(...TRUNCATED)
[ false, false, false, false ]
[ "", "", "", "" ]
{"api_url":"http://10.32.36.236:9080/v1","backend":"litellm","chat_template_applied":true,"generatio(...TRUNCATED)
{"backend":"litellm","model_name":"hosted_vllm/TAUR-dev/M-1119_newmodels__qwen7b_sft1k_ours-rl@check(...TRUNCATED)
[ false, true, false, true ]
[ "idea", "back", "bake", "back" ]
"[{\"extraction_type\": \"final\", \"total_spans_found\": 9, \"non_overlapping_spans\": 9, \"span_st(...TRUNCATED)
"[{\"is_correct\": false, \"error\": \"Constraint violations: violates_character_position_constraint(...TRUNCATED)
[[false,false,false,false,false,false,false,false,false],[true],[false,false,false,false],[false,tru(...TRUNCATED)
[["bach","bake","blie","idea","blade","idea","akis","acke","idea"],["back"],["bake","bake","bake","b(...TRUNCATED)
"[[{\"pattern_used\": \"answer tag\", \"confidence\": 1.0, \"position\": 1868, \"span_text\": \"bach(...TRUNCATED)
"[[{\"is_correct\": false, \"error\": \"Constraint violations: violates_character_position_constrain(...TRUNCATED)
{"flips_by":[0,0,0,1],"flips_total":1,"num_correct":2,"pass_at_n":1,"percent_correct":0.5,"total_res(...TRUNCATED)
2025-11-27
test
main
TAUR-dev/M-1119_newmodels__qwen7b_sft1k_ours-rl@checkpoint-step-150
0
eval_rl
1
2025-11-27T21:20:34.975885
"TAUR-dev/D-EVAL__standard_eval_v3__FinEval_16k_fulleval_3arg_OT_ours_1k-at150-RL-acronym_4o-eval_rl(...TRUNCATED)
"Answer the following problem. Explain your reasoning step by step. When you are finished, give your(...TRUNCATED)
acronym_4o
acronym
[{"content":"Answer the following problem. Explain your reasoning step by step. When you are finishe(...TRUNCATED)
[]
[]
"{\"acronym\": \"life\", \"words\": [\"layer\", \"immigrant\", \"fisher\", \"exhibits\"], \"word_cou(...TRUNCATED)
test
life
life
4
[ "layer", "immigrant", "fisher", "exhibits" ]
["<think>\n<sample>\nTo construct an acronym that meets the given conditions, we need to follow thes(...TRUNCATED)
[ false, false, false, false ]
[ "", "", "", "" ]
{"api_url":"http://10.32.36.236:9080/v1","backend":"litellm","chat_template_applied":true,"generatio(...TRUNCATED)
{"backend":"litellm","model_name":"hosted_vllm/TAUR-dev/M-1119_newmodels__qwen7b_sft1k_ours-rl@check(...TRUNCATED)
[ false, true, true, true ]
[ "live", "life", "life", "life" ]
"[{\"extraction_type\": \"final\", \"total_spans_found\": 5, \"non_overlapping_spans\": 5, \"span_st(...TRUNCATED)
"[{\"is_correct\": false, \"error\": \"Constraint violations: violates_character_position_constraint(...TRUNCATED)
[[false,true,false,false,false],[true,false,true,true,true,true],[true,true,false,true,true],[true,t(...TRUNCATED)
[["list","life","live","live","live"],["life","lift","life","LIFE","life","life"],["life","life","li(...TRUNCATED)
"[[{\"pattern_used\": \"answer tag\", \"confidence\": 1.0, \"position\": 1185, \"span_text\": \"list(...TRUNCATED)
"[[{\"is_correct\": false, \"error\": \"Constraint violations: violates_character_position_constrain(...TRUNCATED)
{"flips_by":[1,1,1,0],"flips_total":3,"num_correct":3,"pass_at_n":1,"percent_correct":0.75,"total_re(...TRUNCATED)
2025-11-27
test
main
TAUR-dev/M-1119_newmodels__qwen7b_sft1k_ours-rl@checkpoint-step-150
0
eval_rl
1
2025-11-27T21:20:34.975888
"TAUR-dev/D-EVAL__standard_eval_v3__FinEval_16k_fulleval_3arg_OT_ours_1k-at150-RL-acronym_4o-eval_rl(...TRUNCATED)
"Answer the following problem. Explain your reasoning step by step. When you are finished, give your(...TRUNCATED)
acronym_4o
acronym
[{"content":"Answer the following problem. Explain your reasoning step by step. When you are finishe(...TRUNCATED)
[]
[]
"{\"acronym\": \"best\", \"words\": [\"bonus\", \"eight\", \"shanghai\", \"them\"], \"word_count\": (...TRUNCATED)
test
best
best
4
[ "bonus", "eight", "shanghai", "them" ]
["<think>\n<sample>\nTo construct an acronym that meets all the given conditions, we need to follow (...TRUNCATED)
[ false, false, false, false ]
[ "", "", "", "" ]
{"api_url":"http://10.32.36.236:9080/v1","backend":"litellm","chat_template_applied":true,"generatio(...TRUNCATED)
{"backend":"litellm","model_name":"hosted_vllm/TAUR-dev/M-1119_newmodels__qwen7b_sft1k_ours-rl@check(...TRUNCATED)
[ true, true, false, true ]
[ "best", "best", "east", "best" ]
"[{\"extraction_type\": \"final\", \"total_spans_found\": 3, \"non_overlapping_spans\": 3, \"span_st(...TRUNCATED)
"[{\"is_correct\": true, \"error\": \"\", \"answer_block\": \"best\", \"final_answer\": \"best\", \"(...TRUNCATED)
[ [ true, true, true ], [ true, true, true, true, true ], [ false, false, false ], [ true, true, true ] ]
[["best","best","best"],["best","best","best","best","best"],["east","east","east"],["best","best","(...TRUNCATED)
"[[{\"pattern_used\": \"answer tag\", \"confidence\": 1.0, \"position\": 1102, \"span_text\": \"best(...TRUNCATED)
"[[{\"is_correct\": true, \"error\": \"\", \"answer_block\": \"best\", \"final_answer\": \"best\", \(...TRUNCATED)
{"flips_by":[0,0,0,0],"flips_total":0,"num_correct":3,"pass_at_n":1,"percent_correct":0.75,"total_re(...TRUNCATED)
2025-11-27
test
main
TAUR-dev/M-1119_newmodels__qwen7b_sft1k_ours-rl@checkpoint-step-150
0
eval_rl
1
2025-11-27T21:20:34.975891
"TAUR-dev/D-EVAL__standard_eval_v3__FinEval_16k_fulleval_3arg_OT_ours_1k-at150-RL-acronym_4o-eval_rl(...TRUNCATED)
"Answer the following problem. Explain your reasoning step by step. When you are finished, give your(...TRUNCATED)
acronym_4o
acronym
[{"content":"Answer the following problem. Explain your reasoning step by step. When you are finishe(...TRUNCATED)
[]
[]
"{\"acronym\": \"long\", \"words\": [\"listen\", \"offensive\", \"nets\", \"galaxy\"], \"word_count\(...TRUNCATED)
test
long
long
4
[ "listen", "offensive", "nets", "galaxy" ]
["<think>\n<sample>\nTo construct an acronym that meets the given conditions, we need to follow thes(...TRUNCATED)
[ false, false, false, false ]
[ "", "", "", "" ]
{"api_url":"http://10.32.36.236:9080/v1","backend":"litellm","chat_template_applied":true,"generatio(...TRUNCATED)
{"backend":"litellm","model_name":"hosted_vllm/TAUR-dev/M-1119_newmodels__qwen7b_sft1k_ours-rl@check(...TRUNCATED)
[ true, false, true, true ]
[ "long", "lone", "long", "long" ]
"[{\"extraction_type\": \"final\", \"total_spans_found\": 6, \"non_overlapping_spans\": 6, \"span_st(...TRUNCATED)
"[{\"is_correct\": true, \"error\": \"\", \"answer_block\": \"long\", \"final_answer\": \"long\", \"(...TRUNCATED)
[[true,true,true,false,true,true],[false,true,false,false,false,false,false],[false,false,true,true,(...TRUNCATED)
[["long","LONG","LONG","logy","long","long"],["lone","long","lots","loin","loan","lone","lone"],["lo(...TRUNCATED)
"[[{\"pattern_used\": \"answer tag\", \"confidence\": 1.0, \"position\": 1241, \"span_text\": \"long(...TRUNCATED)
"[[{\"is_correct\": true, \"error\": \"\", \"answer_block\": \"long\", \"final_answer\": \"long\", \(...TRUNCATED)
{"flips_by":[1,1,0,0],"flips_total":2,"num_correct":3,"pass_at_n":1,"percent_correct":0.75,"total_re(...TRUNCATED)
2025-11-27
test
main
TAUR-dev/M-1119_newmodels__qwen7b_sft1k_ours-rl@checkpoint-step-150
0
eval_rl
1
2025-11-27T21:20:34.975892
"TAUR-dev/D-EVAL__standard_eval_v3__FinEval_16k_fulleval_3arg_OT_ours_1k-at150-RL-acronym_4o-eval_rl(...TRUNCATED)
"Answer the following problem. Explain your reasoning step by step. When you are finished, give your(...TRUNCATED)
acronym_4o
acronym
[{"content":"Answer the following problem. Explain your reasoning step by step. When you are finishe(...TRUNCATED)
[]
[]
"{\"acronym\": \"part\", \"words\": [\"proposed\", \"agreements\", \"randomly\", \"tube\"], \"word_c(...TRUNCATED)
test
part
part
4
[ "proposed", "agreements", "randomly", "tube" ]
["<think>\n<sample>\nTo construct an acronym that meets all the given conditions, we need to follow (...TRUNCATED)
[ false, false, false, false ]
[ "", "", "", "" ]
{"api_url":"http://10.32.36.236:9080/v1","backend":"litellm","chat_template_applied":true,"generatio(...TRUNCATED)
{"backend":"litellm","model_name":"hosted_vllm/TAUR-dev/M-1119_newmodels__qwen7b_sft1k_ours-rl@check(...TRUNCATED)
[ false, true, true, true ]
[ "pat", "part", "part", "part" ]
"[{\"extraction_type\": \"final\", \"total_spans_found\": 28, \"non_overlapping_spans\": 28, \"span_(...TRUNCATED)
"[{\"is_correct\": false, \"error\": \"Constraint violations: invalid_length\", \"answer_block\": \"(...TRUNCATED)
[[false,false,true,false,false,true,true,true,false,false,false,false,false,false,false,false,false,(...TRUNCATED)
[["pact","pact","part","tap","pat","part","part","part","pat","pat","pat","pat","pat","pat","pat","p(...TRUNCATED)
"[[{\"pattern_used\": \"answer tag\", \"confidence\": 1.0, \"position\": 1288, \"span_text\": \"pact(...TRUNCATED)
"[[{\"is_correct\": false, \"error\": \"Constraint violations: violates_character_position_constrain(...TRUNCATED)
{"flips_by":[2,0,1,0],"flips_total":3,"num_correct":3,"pass_at_n":1,"percent_correct":0.75,"total_re(...TRUNCATED)
2025-11-27
test
main
TAUR-dev/M-1119_newmodels__qwen7b_sft1k_ours-rl@checkpoint-step-150
0
eval_rl
1
2025-11-27T21:20:34.975894
"TAUR-dev/D-EVAL__standard_eval_v3__FinEval_16k_fulleval_3arg_OT_ours_1k-at150-RL-acronym_4o-eval_rl(...TRUNCATED)
"Answer the following problem. Explain your reasoning step by step. When you are finished, give your(...TRUNCATED)
acronym_4o
acronym
[{"content":"Answer the following problem. Explain your reasoning step by step. When you are finishe(...TRUNCATED)
[]
[]
"{\"acronym\": \"find\", \"words\": [\"folks\", \"investigated\", \"next\", \"destinations\"], \"wor(...TRUNCATED)
test
find
find
4
[ "folks", "investigated", "next", "destinations" ]
["<think>\n<sample>\nTo construct an acronym that meets all the given conditions, we need to follow (...TRUNCATED)
[ false, false, false, false ]
[ "", "", "", "" ]
{"api_url":"http://10.32.36.236:9080/v1","backend":"litellm","chat_template_applied":true,"generatio(...TRUNCATED)
{"backend":"litellm","model_name":"hosted_vllm/TAUR-dev/M-1119_newmodels__qwen7b_sft1k_ours-rl@check(...TRUNCATED)
[ true, true, true, true ]
[ "find", "find", "find", "find" ]
"[{\"extraction_type\": \"final\", \"total_spans_found\": 4, \"non_overlapping_spans\": 4, \"span_st(...TRUNCATED)
"[{\"is_correct\": true, \"error\": \"\", \"answer_block\": \"find\", \"final_answer\": \"find\", \"(...TRUNCATED)
[ [ true, false, true, true ], [ true, true, true, true, true ], [ true, true, true ], [ true, true, true, true, true ] ]
[["find","finest","find","find"],["find","FIND","find","find","find"],["find","find","find"],["find"(...TRUNCATED)
"[[{\"pattern_used\": \"answer tag\", \"confidence\": 1.0, \"position\": 865, \"span_text\": \"find\(...TRUNCATED)
"[[{\"is_correct\": true, \"error\": \"\", \"answer_block\": \"find\", \"final_answer\": \"find\", \(...TRUNCATED)
{"flips_by":[1,0,0,0],"flips_total":1,"num_correct":4,"pass_at_n":1,"percent_correct":1.0,"total_res(...TRUNCATED)
2025-11-27
test
main
TAUR-dev/M-1119_newmodels__qwen7b_sft1k_ours-rl@checkpoint-step-150
0
eval_rl
1
2025-11-27T21:20:34.975896
"TAUR-dev/D-EVAL__standard_eval_v3__FinEval_16k_fulleval_3arg_OT_ours_1k-at150-RL-acronym_4o-eval_rl(...TRUNCATED)
"Answer the following problem. Explain your reasoning step by step. When you are finished, give your(...TRUNCATED)
acronym_4o
acronym
[{"content":"Answer the following problem. Explain your reasoning step by step. When you are finishe(...TRUNCATED)
[]
[]
"{\"acronym\": \"does\", \"words\": [\"donated\", \"organisations\", \"efforts\", \"scientists\"], \(...TRUNCATED)
test
does
does
4
[ "donated", "organisations", "efforts", "scientists" ]
["<think>\n<sample>\nTo construct an acronym that meets the given conditions, we need to follow thes(...TRUNCATED)
[ false, false, false, false ]
[ "", "", "", "" ]
{"api_url":"http://10.32.36.236:9080/v1","backend":"litellm","chat_template_applied":true,"generatio(...TRUNCATED)
{"backend":"litellm","model_name":"hosted_vllm/TAUR-dev/M-1119_newmodels__qwen7b_sft1k_ours-rl@check(...TRUNCATED)
[ true, true, true, false ]
[ "does", "does", "does", "dots" ]
"[{\"extraction_type\": \"final\", \"total_spans_found\": 5, \"non_overlapping_spans\": 5, \"span_st(...TRUNCATED)
"[{\"is_correct\": true, \"error\": \"\", \"answer_block\": \"does\", \"final_answer\": \"does\", \"(...TRUNCATED)
[[false,false,true,true,true],[false,false,true],[true,false,true,true,true],[false,false,true,false(...TRUNCATED)
[["dose","des","does","does","does"],["doe","odes","does"],["does","dens","does","does","does"],["do(...TRUNCATED)
"[[{\"pattern_used\": \"answer tag\", \"confidence\": 1.0, \"position\": 1332, \"span_text\": \"dose(...TRUNCATED)
"[[{\"is_correct\": false, \"error\": \"Constraint violations: violates_character_position_constrain(...TRUNCATED)
{"flips_by":[0,0,1,1],"flips_total":2,"num_correct":3,"pass_at_n":1,"percent_correct":0.75,"total_re(...TRUNCATED)
2025-11-27
test
main
TAUR-dev/M-1119_newmodels__qwen7b_sft1k_ours-rl@checkpoint-step-150
0
eval_rl
1
2025-11-27T21:20:34.975898
"TAUR-dev/D-EVAL__standard_eval_v3__FinEval_16k_fulleval_3arg_OT_ours_1k-at150-RL-acronym_4o-eval_rl(...TRUNCATED)
End of preview. Expand in Data Studio

Experiment Tracker: FinEval_16k_fulleval_3arg_OT_ours_1k-at150-RL-acronym_4o

Experiment Description: Evaluation experiment for task acronym_4o from FinEval_16k_fulleval_3arg_OT_ours_1k-at150-RL

Start Time: 2025-11-27T20:45:26.385668

Tracker Dataset: TAUR-dev/D-ExpTracker__FinEval_16k_fulleval_3arg_OT_ours_1k-at150-RL-acronym_4o__v1

Stages Completed

Total stages: 1

Models Created

Dataset Configurations

This tracker dataset contains the following configurations with immediate upload as stages complete:

Training Data (Complete Datasets)

Hyperparameters (Complete Configurations)

Logs (Stage-Specific)

Evaluation Results (Complete with Annotations)

Metadata

  • experiment_metadata: Timeline and stage information

Usage

Load specific configurations with:

from datasets import load_dataset

# Load experiment metadata
metadata = load_dataset('TAUR-dev/D-ExpTracker__FinEval_16k_fulleval_3arg_OT_ours_1k-at150-RL-acronym_4o__v1', 'experiment_metadata')

# Load complete training datasets
sft_data = load_dataset('TAUR-dev/D-ExpTracker__FinEval_16k_fulleval_3arg_OT_ours_1k-at150-RL-acronym_4o__v1', 'training_data__sft')
sft_metadata = load_dataset('TAUR-dev/D-ExpTracker__FinEval_16k_fulleval_3arg_OT_ours_1k-at150-RL-acronym_4o__v1', 'training_data__sft_metadata')

# Load complete configurations
sft_hyperparams = load_dataset('TAUR-dev/D-ExpTracker__FinEval_16k_fulleval_3arg_OT_ours_1k-at150-RL-acronym_4o__v1', 'hyperparameters__sft')
rl_hyperparams = load_dataset('TAUR-dev/D-ExpTracker__FinEval_16k_fulleval_3arg_OT_ours_1k-at150-RL-acronym_4o__v1', 'hyperparameters__rl')

# Load stage-specific logs
sft_logs = load_dataset('TAUR-dev/D-ExpTracker__FinEval_16k_fulleval_3arg_OT_ours_1k-at150-RL-acronym_4o__v1', 'logs__sft')
rl_logs = load_dataset('TAUR-dev/D-ExpTracker__FinEval_16k_fulleval_3arg_OT_ours_1k-at150-RL-acronym_4o__v1', 'logs__rl')

# Load evaluation results with annotations
sft_eval_results = load_dataset('TAUR-dev/D-ExpTracker__FinEval_16k_fulleval_3arg_OT_ours_1k-at150-RL-acronym_4o__v1', 'evals_eval_sft')
rl_eval_results = load_dataset('TAUR-dev/D-ExpTracker__FinEval_16k_fulleval_3arg_OT_ours_1k-at150-RL-acronym_4o__v1', 'evals_eval_rl')

Models

Registry

All models from this experiment are automatically registered in the SkillFactory Model Registry with:

  • Complete training configuration (hyperparameters, datasets, methods)
  • Experiment lineage (links back to this tracker dataset)
  • Stage-specific metadata (SFT vs RL training details)
  • Structured input data references (training datasets and configurations)

Registry entries follow the naming pattern: Model - FinEval_16k_fulleval_3arg_OT_ours_1k-at150-RL-acronym_4o - {stage_name} - {SFT/RL}


Generated by SkillFactory Experiment Management System All artifacts uploaded immediately as stages complete with perfect data provenance

Downloads last month
4