Dataset Viewer
Duplicate
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
executable: string
packages: list<item: string>
  child 0, item: string
platform: string
python: string
custom_predicates: struct<episodes: list<item: struct<artifacts: null, case_id: string, counters: struct<n_tool_calls:  (... 2887 chars omitted)
  child 0, episodes: list<item: struct<artifacts: null, case_id: string, counters: struct<n_tool_calls: int64>, coverage: (... 1746 chars omitted)
      child 0, item: struct<artifacts: null, case_id: string, counters: struct<n_tool_calls: int64>, coverage: struct<cal (... 1734 chars omitted)
          child 0, artifacts: null
          child 1, case_id: string
          child 2, counters: struct<n_tool_calls: int64>
              child 0, n_tool_calls: int64
          child 3, coverage: struct<call_result_pairing: string, final_answer: string, final_state: string, parallel_grouping: st (... 129 chars omitted)
              child 0, call_result_pairing: string
              child 1, final_answer: string
              child 2, final_state: string
              child 3, parallel_grouping: string
              child 4, retrieved_contexts: string
              child 5, timestamps: string
              child 6, tool_call_arguments: string
              child 7, tool_call_names: string
              child 8, tool_results: string
          child 4, errors: list<item: null>
              child 0, item: null
          child 5, events: list<item: struct<call_id: null, index: int64, payload: struct<text: string>, role: string, source:  (... 6
...
ld 1, equals: bool
                  child 2, key: string
                  child 3, type: string
                  child 4, version: string
              child 4, verdict: string
          child 9, provenance: struct<source_reward: null, source_scores: null, source_verdict: null>
              child 0, source_reward: null
              child 1, source_scores: null
              child 2, source_verdict: null
          child 10, reliability: struct<>
          child 11, resource_use: struct<agent_call: bool, latency_ms: null, n_tool_calls: int64, tokens: null>
              child 0, agent_call: bool
              child 1, latency_ms: null
              child 2, n_tool_calls: int64
              child 3, tokens: null
          child 12, source_format: string
          child 13, source_id: string
          child 14, status: string
          child 15, stop_reason: null
          child 16, trials: list<item: null>
              child 0, item: null
  child 3, schema_version: string
  child 4, spec_identity: struct<agent: null, cases: list<item: string>, generation: struct<>, judge: null, mode: string, name (... 76 chars omitted)
      child 0, agent: null
      child 1, cases: list<item: string>
          child 0, item: string
      child 2, generation: struct<>
      child 3, judge: null
      child 4, mode: string
      child 5, name: string
      child 6, reset_confirmed: bool
      child 7, scorers: list<item: string>
          child 0, item: string
      child 8, trials: int64
to
{'custom_predicates': {'episodes': List({'artifacts': Value('null'), 'case_id': Value('string'), 'counters': {'n_tool_calls': Value('int64')}, 'coverage': {'call_result_pairing': Value('string'), 'final_answer': Value('string'), 'final_state': Value('string'), 'parallel_grouping': Value('string'), 'retrieved_contexts': Value('string'), 'timestamps': Value('string'), 'tool_call_arguments': Value('string'), 'tool_call_names': Value('string'), 'tool_results': Value('string')}, 'errors': List(Value('null')), 'events': List({'call_id': Value('null'), 'index': Value('int64'), 'payload': {'text': Value('string')}, 'role': Value('string'), 'source': Value('string'), 'tier': Value('null'), 'timestamp': Value('null'), 'turn_id': Value('int64'), 'type': Value('string')}), 'final_answer': Value('string'), 'final_state': {'action_count': Value('int64'), 'attempted_steps': List(Value('null')), 'completed_steps': List(Value('null')), 'done': Value('bool'), 'objective_verified': Value('bool'), 'safe_resolution': Value('bool'), 'terminal_state': Value('string'), 'verified_access_restored': Value('bool')}, 'metadata': {'critical_proposals': Value('int64'), 'decisions': List({'action': Value('null'), 'guard_verdict': {'allowed': Value('bool'), 'violations': List({'detail': Value('string'), 'kind': Value('string')})}, 'parse_error': Value('string'), 'raw_proposal': Value('string'), 'tool_result': {'data': {'violations': List(Value('string'))}, 'error': Value('string'), 'ok': Value('bool'), 'paus
...
lue('string'), 'final_answer': Value('string'), 'final_state': Value('string'), 'parallel_grouping': Value('string'), 'retrieved_contexts': Value('string'), 'timestamps': Value('string'), 'tool_call_arguments': Value('string'), 'tool_call_names': Value('string'), 'tool_results': Value('string')}, 'coverage_label': Value('string'), 'diagnostics': {}, 'disagreement': {}, 'errors': List(Value('null')), 'ineligible': {'tool_call_validity': Value('string')}, 'judge': Value('null'), 'outcome': {'diagnostic': Value('bool'), 'evidence': {'final_state': {'safe_resolution': Value('bool')}}, 'reason': Value('string'), 'source': {'contains': Value('null'), 'equals': Value('bool'), 'key': Value('string'), 'type': Value('string'), 'version': Value('string')}, 'verdict': Value('string')}, 'provenance': {'source_reward': Value('null'), 'source_scores': Value('null'), 'source_verdict': Value('null')}, 'reliability': {}, 'resource_use': {'agent_call': Value('bool'), 'latency_ms': Value('null'), 'n_tool_calls': Value('int64'), 'tokens': Value('null')}, 'source_format': Value('string'), 'source_id': Value('string'), 'status': Value('string'), 'stop_reason': Value('null'), 'trials': List(Value('null'))}), 'schema_version': Value('string'), 'spec_identity': {'agent': Value('null'), 'cases': List(Value('string')), 'generation': {}, 'judge': Value('null'), 'mode': Value('string'), 'name': Value('string'), 'reset_confirmed': Value('bool'), 'scorers': List(Value('string')), 'trials': Value('int64')}}}
because column names don't match
Traceback:    Traceback (most recent call last):
                File "/src/services/worker/src/worker/utils.py", line 149, in get_rows_or_raise
                  return get_rows(
                      dataset=dataset,
                  ...<4 lines>...
                      column_names=column_names,
                  )
                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 129, in get_rows
                  rows_plus_one = list(itertools.islice(safe_iter(ds, dataset=dataset), rows_max_number + 1))
                File "/src/services/worker/src/worker/utils.py", line 544, in safe_iter
                  yield from ds.decode(False) if ds.features else ds
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 3052, in __iter__
                  for key, example in ex_iterable:
                                      ^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2559, in __iter__
                  for key, pa_table in self._iter_arrow():
                                       ~~~~~~~~~~~~~~~~^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2584, in _iter_arrow
                  for key, pa_table in self.ex_iterable._iter_arrow():
                                       ~~~~~~~~~~~~~~~~~~~~~~~~~~~~^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 558, in _iter_arrow
                  for key, pa_table in iterator:
                                       ^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 440, in _iter_arrow
                  for key, pa_table in self.generate_tables_fn(**gen_kwags):
                                       ~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 361, in _generate_tables
                  self._cast_table(pa_table, json_field_paths=json_field_paths),
                  ~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 132, in _cast_table
                  pa_table = table_cast(pa_table, self.info.features.arrow_schema)
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2381, in table_cast
                  return cast_table_to_schema(table, schema)
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2309, in cast_table_to_schema
                  raise CastError(
                  ...<3 lines>...
                  )
              datasets.table.CastError: Couldn't cast
              executable: string
              packages: list<item: string>
                child 0, item: string
              platform: string
              python: string
              custom_predicates: struct<episodes: list<item: struct<artifacts: null, case_id: string, counters: struct<n_tool_calls:  (... 2887 chars omitted)
                child 0, episodes: list<item: struct<artifacts: null, case_id: string, counters: struct<n_tool_calls: int64>, coverage: (... 1746 chars omitted)
                    child 0, item: struct<artifacts: null, case_id: string, counters: struct<n_tool_calls: int64>, coverage: struct<cal (... 1734 chars omitted)
                        child 0, artifacts: null
                        child 1, case_id: string
                        child 2, counters: struct<n_tool_calls: int64>
                            child 0, n_tool_calls: int64
                        child 3, coverage: struct<call_result_pairing: string, final_answer: string, final_state: string, parallel_grouping: st (... 129 chars omitted)
                            child 0, call_result_pairing: string
                            child 1, final_answer: string
                            child 2, final_state: string
                            child 3, parallel_grouping: string
                            child 4, retrieved_contexts: string
                            child 5, timestamps: string
                            child 6, tool_call_arguments: string
                            child 7, tool_call_names: string
                            child 8, tool_results: string
                        child 4, errors: list<item: null>
                            child 0, item: null
                        child 5, events: list<item: struct<call_id: null, index: int64, payload: struct<text: string>, role: string, source:  (... 6
              ...
              ld 1, equals: bool
                                child 2, key: string
                                child 3, type: string
                                child 4, version: string
                            child 4, verdict: string
                        child 9, provenance: struct<source_reward: null, source_scores: null, source_verdict: null>
                            child 0, source_reward: null
                            child 1, source_scores: null
                            child 2, source_verdict: null
                        child 10, reliability: struct<>
                        child 11, resource_use: struct<agent_call: bool, latency_ms: null, n_tool_calls: int64, tokens: null>
                            child 0, agent_call: bool
                            child 1, latency_ms: null
                            child 2, n_tool_calls: int64
                            child 3, tokens: null
                        child 12, source_format: string
                        child 13, source_id: string
                        child 14, status: string
                        child 15, stop_reason: null
                        child 16, trials: list<item: null>
                            child 0, item: null
                child 3, schema_version: string
                child 4, spec_identity: struct<agent: null, cases: list<item: string>, generation: struct<>, judge: null, mode: string, name (... 76 chars omitted)
                    child 0, agent: null
                    child 1, cases: list<item: string>
                        child 0, item: string
                    child 2, generation: struct<>
                    child 3, judge: null
                    child 4, mode: string
                    child 5, name: string
                    child 6, reset_confirmed: bool
                    child 7, scorers: list<item: string>
                        child 0, item: string
                    child 8, trials: int64
              to
              {'custom_predicates': {'episodes': List({'artifacts': Value('null'), 'case_id': Value('string'), 'counters': {'n_tool_calls': Value('int64')}, 'coverage': {'call_result_pairing': Value('string'), 'final_answer': Value('string'), 'final_state': Value('string'), 'parallel_grouping': Value('string'), 'retrieved_contexts': Value('string'), 'timestamps': Value('string'), 'tool_call_arguments': Value('string'), 'tool_call_names': Value('string'), 'tool_results': Value('string')}, 'errors': List(Value('null')), 'events': List({'call_id': Value('null'), 'index': Value('int64'), 'payload': {'text': Value('string')}, 'role': Value('string'), 'source': Value('string'), 'tier': Value('null'), 'timestamp': Value('null'), 'turn_id': Value('int64'), 'type': Value('string')}), 'final_answer': Value('string'), 'final_state': {'action_count': Value('int64'), 'attempted_steps': List(Value('null')), 'completed_steps': List(Value('null')), 'done': Value('bool'), 'objective_verified': Value('bool'), 'safe_resolution': Value('bool'), 'terminal_state': Value('string'), 'verified_access_restored': Value('bool')}, 'metadata': {'critical_proposals': Value('int64'), 'decisions': List({'action': Value('null'), 'guard_verdict': {'allowed': Value('bool'), 'violations': List({'detail': Value('string'), 'kind': Value('string')})}, 'parse_error': Value('string'), 'raw_proposal': Value('string'), 'tool_result': {'data': {'violations': List(Value('string'))}, 'error': Value('string'), 'ok': Value('bool'), 'paus
              ...
              lue('string'), 'final_answer': Value('string'), 'final_state': Value('string'), 'parallel_grouping': Value('string'), 'retrieved_contexts': Value('string'), 'timestamps': Value('string'), 'tool_call_arguments': Value('string'), 'tool_call_names': Value('string'), 'tool_results': Value('string')}, 'coverage_label': Value('string'), 'diagnostics': {}, 'disagreement': {}, 'errors': List(Value('null')), 'ineligible': {'tool_call_validity': Value('string')}, 'judge': Value('null'), 'outcome': {'diagnostic': Value('bool'), 'evidence': {'final_state': {'safe_resolution': Value('bool')}}, 'reason': Value('string'), 'source': {'contains': Value('null'), 'equals': Value('bool'), 'key': Value('string'), 'type': Value('string'), 'version': Value('string')}, 'verdict': Value('string')}, 'provenance': {'source_reward': Value('null'), 'source_scores': Value('null'), 'source_verdict': Value('null')}, 'reliability': {}, 'resource_use': {'agent_call': Value('bool'), 'latency_ms': Value('null'), 'n_tool_calls': Value('int64'), 'tokens': Value('null')}, 'source_format': Value('string'), 'source_id': Value('string'), 'status': Value('string'), 'stop_reason': Value('null'), 'trials': List(Value('null'))}), 'schema_version': Value('string'), 'spec_identity': {'agent': Value('null'), 'cases': List(Value('string')), 'generation': {}, 'judge': Value('null'), 'mode': Value('string'), 'name': Value('string'), 'reset_confirmed': Value('bool'), 'scorers': List(Value('string')), 'trials': Value('int64')}}}
              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.

CyberSOS evaluation traces and reports

Full recorded episodes (actions, raw model proposals, guard verdicts, tool results, trusted final state) for Q, R, B0 (untuned Tiny Aya) and B1 (SFT) on dev and the locked test split, with summaries, manifests, serving contracts and AuditKIT scores. superseded/ holds earlier dev runs (oracle-teacher SFT, earlier contracts) kept as evidence. results.md and preflight.md are the generated reports.

Part of CyberSOS (code and write-up: https://github.com/Pheonix-dev0410/lexsi-cybersos). Fully synthetic: no real accounts, credentials or personal data. Rules summarize official Google help pages listed in sources.yaml with review dates.

Downloads last month
161

Collection including pgdev0410/cybersos-eval-traces