Datasets:

mindbomber
/

aana-tau2-bench-gpt41mini-1trial

Tags:

License:

Dataset card Files Files and versions

xet

Community

Dataset Viewer

Duplicate

Split (1)

train

The dataset viewer is not available for this split.

Cannot load the dataset split (in streaming mode) to extract the first rows.

Error code:   StreamingRowsError
Exception:    CastError
Message:      Couldn't cast
timestamp: string
info: struct<git_commit: string, num_trials: int64, max_steps: int64, max_errors: int64, user_info: struct (... 545 chars omitted)
  child 0, git_commit: string
  child 1, num_trials: int64
  child 2, max_steps: int64
  child 3, max_errors: int64
  child 4, user_info: struct<implementation: string, llm: string, llm_args: struct<temperature: double>, global_simulation (... 64 chars omitted)
      child 0, implementation: string
      child 1, llm: string
      child 2, llm_args: struct<temperature: double>
          child 0, temperature: double
      child 3, global_simulation_guidelines: string
      child 4, voice_settings: null
      child 5, persona_config: null
  child 5, agent_info: struct<implementation: string, llm: string, llm_args: struct<temperature: double>, voice_settings: n (... 4 chars omitted)
      child 0, implementation: string
      child 1, llm: string
      child 2, llm_args: struct<temperature: double>
          child 0, temperature: double
      child 3, voice_settings: null
  child 6, environment_info: struct<domain_name: string, policy: string, tool_defs: null>
      child 0, domain_name: string
      child 1, policy: string
      child 2, tool_defs: null
  child 7, seed: int64
  child 8, text_streaming_config: struct<chunk_by: string, chunk_size: int64>
      child 0, chunk_by: string
      child 1, chunk_size: int64
  child 9, speech_complexity: null
  child 10, audio_native_config: null
  child 11, retrieval_config: null
  child 1
...
1, omitted_questions: bool
      child 2, details: string
contact_info: struct<email: string, name: string, github: string>
  child 0, email: string
  child 1, name: string
  child 2, github: string
results: struct<retail: struct<pass_1: double, cost: double>, airline: struct<pass_1: double, cost: double>,  (... 129 chars omitted)
  child 0, retail: struct<pass_1: double, cost: double>
      child 0, pass_1: double
      child 1, cost: double
  child 1, airline: struct<pass_1: double, cost: double>
      child 0, pass_1: double
      child 1, cost: double
  child 2, telecom: struct<pass_1: double, cost: double>
      child 0, pass_1: double
      child 1, cost: double
  child 3, banking_knowledge: struct<pass_1: double, cost: double, retrieval_config: string>
      child 0, pass_1: double
      child 1, cost: double
      child 2, retrieval_config: string
submitting_organization: string
submission_date: timestamp[s]
trajectories_available: bool
submission_type: string
model_name: string
modality: string
trajectory_files: struct<airline: string, banking_knowledge: string, retail: string, telecom: string>
  child 0, airline: string
  child 1, banking_knowledge: string
  child 2, retail: string
  child 3, telecom: string
references: list<item: struct<title: string, url: string, type: string>>
  child 0, item: struct<title: string, url: string, type: string>
      child 0, title: string
      child 1, url: string
      child 2, type: string
is_new: bool
model_organization: string
to
{'model_name': Value('string'), 'model_organization': Value('string'), 'submitting_organization': Value('string'), 'submission_date': Value('timestamp[s]'), 'submission_type': Value('string'), 'modality': Value('string'), 'contact_info': {'email': Value('string'), 'name': Value('string'), 'github': Value('string')}, 'results': {'retail': {'pass_1': Value('float64'), 'cost': Value('float64')}, 'airline': {'pass_1': Value('float64'), 'cost': Value('float64')}, 'telecom': {'pass_1': Value('float64'), 'cost': Value('float64')}, 'banking_knowledge': {'pass_1': Value('float64'), 'cost': Value('float64'), 'retrieval_config': Value('string')}}, 'is_new': Value('bool'), 'trajectories_available': Value('bool'), 'trajectory_files': {'airline': Value('string'), 'banking_knowledge': Value('string'), 'retail': Value('string'), 'telecom': Value('string')}, 'references': List({'title': Value('string'), 'url': Value('string'), 'type': Value('string')}), 'methodology': {'evaluation_date': Value('timestamp[s]'), 'tau2_bench_version': Value('string'), 'user_simulator': Value('string'), 'notes': Value('string'), 'verification': {'modified_prompts': Value('bool'), 'omitted_questions': Value('bool'), 'details': Value('string')}}}
because column names don't match
Traceback:    Traceback (most recent call last):
                File "/src/services/worker/src/worker/utils.py", line 99, in get_rows_or_raise
                  return get_rows(
                         ^^^^^^^^^
                File "/src/libs/libcommon/src/libcommon/utils.py", line 272, in decorator
                  return func(*args, **kwargs)
                         ^^^^^^^^^^^^^^^^^^^^^
                File "/src/services/worker/src/worker/utils.py", line 77, in get_rows
                  rows_plus_one = list(itertools.islice(ds, rows_max_number + 1))
                                  ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/iterable_dataset.py", line 2690, in __iter__
                  for key, example in ex_iterable:
                                      ^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/iterable_dataset.py", line 2227, in __iter__
                  for key, pa_table in self._iter_arrow():
                                       ^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/iterable_dataset.py", line 2251, in _iter_arrow
                  for key, pa_table in self.ex_iterable._iter_arrow():
                                       ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/iterable_dataset.py", line 494, in _iter_arrow
                  for key, pa_table in iterator:
                                       ^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/iterable_dataset.py", line 384, in _iter_arrow
                  for key, pa_table in self.generate_tables_fn(**gen_kwags):
                                       ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/packaged_modules/json/json.py", line 299, in _generate_tables
                  self._cast_table(pa_table, json_field_paths=json_field_paths),
                  ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/packaged_modules/json/json.py", line 128, in _cast_table
                  pa_table = table_cast(pa_table, self.info.features.arrow_schema)
                             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/table.py", line 2321, in table_cast
                  return cast_table_to_schema(table, schema)
                         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/table.py", line 2249, in cast_table_to_schema
                  raise CastError(
              datasets.table.CastError: Couldn't cast
              timestamp: string
              info: struct<git_commit: string, num_trials: int64, max_steps: int64, max_errors: int64, user_info: struct (... 545 chars omitted)
                child 0, git_commit: string
                child 1, num_trials: int64
                child 2, max_steps: int64
                child 3, max_errors: int64
                child 4, user_info: struct<implementation: string, llm: string, llm_args: struct<temperature: double>, global_simulation (... 64 chars omitted)
                    child 0, implementation: string
                    child 1, llm: string
                    child 2, llm_args: struct<temperature: double>
                        child 0, temperature: double
                    child 3, global_simulation_guidelines: string
                    child 4, voice_settings: null
                    child 5, persona_config: null
                child 5, agent_info: struct<implementation: string, llm: string, llm_args: struct<temperature: double>, voice_settings: n (... 4 chars omitted)
                    child 0, implementation: string
                    child 1, llm: string
                    child 2, llm_args: struct<temperature: double>
                        child 0, temperature: double
                    child 3, voice_settings: null
                child 6, environment_info: struct<domain_name: string, policy: string, tool_defs: null>
                    child 0, domain_name: string
                    child 1, policy: string
                    child 2, tool_defs: null
                child 7, seed: int64
                child 8, text_streaming_config: struct<chunk_by: string, chunk_size: int64>
                    child 0, chunk_by: string
                    child 1, chunk_size: int64
                child 9, speech_complexity: null
                child 10, audio_native_config: null
                child 11, retrieval_config: null
                child 1
              ...
              1, omitted_questions: bool
                    child 2, details: string
              contact_info: struct<email: string, name: string, github: string>
                child 0, email: string
                child 1, name: string
                child 2, github: string
              results: struct<retail: struct<pass_1: double, cost: double>, airline: struct<pass_1: double, cost: double>,  (... 129 chars omitted)
                child 0, retail: struct<pass_1: double, cost: double>
                    child 0, pass_1: double
                    child 1, cost: double
                child 1, airline: struct<pass_1: double, cost: double>
                    child 0, pass_1: double
                    child 1, cost: double
                child 2, telecom: struct<pass_1: double, cost: double>
                    child 0, pass_1: double
                    child 1, cost: double
                child 3, banking_knowledge: struct<pass_1: double, cost: double, retrieval_config: string>
                    child 0, pass_1: double
                    child 1, cost: double
                    child 2, retrieval_config: string
              submitting_organization: string
              submission_date: timestamp[s]
              trajectories_available: bool
              submission_type: string
              model_name: string
              modality: string
              trajectory_files: struct<airline: string, banking_knowledge: string, retail: string, telecom: string>
                child 0, airline: string
                child 1, banking_knowledge: string
                child 2, retail: string
                child 3, telecom: string
              references: list<item: struct<title: string, url: string, type: string>>
                child 0, item: struct<title: string, url: string, type: string>
                    child 0, title: string
                    child 1, url: string
                    child 2, type: string
              is_new: bool
              model_organization: string
              to
              {'model_name': Value('string'), 'model_organization': Value('string'), 'submitting_organization': Value('string'), 'submission_date': Value('timestamp[s]'), 'submission_type': Value('string'), 'modality': Value('string'), 'contact_info': {'email': Value('string'), 'name': Value('string'), 'github': Value('string')}, 'results': {'retail': {'pass_1': Value('float64'), 'cost': Value('float64')}, 'airline': {'pass_1': Value('float64'), 'cost': Value('float64')}, 'telecom': {'pass_1': Value('float64'), 'cost': Value('float64')}, 'banking_knowledge': {'pass_1': Value('float64'), 'cost': Value('float64'), 'retrieval_config': Value('string')}}, 'is_new': Value('bool'), 'trajectories_available': Value('bool'), 'trajectory_files': {'airline': Value('string'), 'banking_knowledge': Value('string'), 'retail': Value('string'), 'telecom': Value('string')}, 'references': List({'title': Value('string'), 'url': Value('string'), 'type': Value('string')}), 'methodology': {'evaluation_date': Value('timestamp[s]'), 'tau2_bench_version': Value('string'), 'user_simulator': Value('string'), 'notes': Value('string'), 'verification': {'modified_prompts': Value('bool'), 'omitted_questions': Value('bool'), 'details': Value('string')}}}
              because column names don't match

Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.

AANA Contract Gate + GPT-4.1-mini τ²-Bench Submission Artifact

This artifact contains the official tau2 submit prepare output for a custom AANA pre-tool-call contract gate wrapped around the τ²-Bench text agent.

Evaluation date: 2026-05-07

Configuration:

Agent model: openai/gpt-4.1-mini
User simulator: openai/gpt-4.1-mini
Trials: 1 per task
Domains: airline, retail, telecom, banking_knowledge
Banking retrieval: bm25
Submission type: custom

Important limitation: this is an architecture validation run, not a 4-trial statistical reliability leaderboard claim. The custom AANA scaffold gates tool calls before execution with accept/ask/defer/refuse decisions; no tasks were omitted.

Prepared and validated with:

tau2 submit prepare ...
tau2 submit validate ...

Public references:

AANA demo Space: https://huggingface.co/spaces/mindbomber/aana-demo
AANA model card: https://huggingface.co/mindbomber/aana
AANA docs: https://mindbomber.github.io/Alignment-Aware-Neural-Architecture--AANA-/

Implementation reference:

implementation/aana_contract_agent.py

Downloads last month: 56

Total file size:

69.4 MB