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
config: string
config_label: string
direction_layer: int64
fingerprint: string
total_units: int64
units_generated: int64
units_resumed: int64
row_count_per_arm: int64
arms: list<item: string>
  child 0, item: string
by_arm: struct<baseline: struct<known_refused: struct<n: int64, refusal_rate: double, correct_rate: double>, (... 635 chars omitted)
  child 0, baseline: struct<known_refused: struct<n: int64, refusal_rate: double, correct_rate: double>, known_correct_an (... 69 chars omitted)
      child 0, known_refused: struct<n: int64, refusal_rate: double, correct_rate: double>
          child 0, n: int64
          child 1, refusal_rate: double
          child 2, correct_rate: double
      child 1, known_correct_answered: struct<n: int64, refusal_rate: double, correct_rate: double>
          child 0, n: int64
          child 1, refusal_rate: double
          child 2, correct_rate: double
  child 1, ablate: struct<known_refused: struct<n: int64, refusal_rate: double, correct_rate: double>, known_correct_an (... 69 chars omitted)
      child 0, known_refused: struct<n: int64, refusal_rate: double, correct_rate: double>
          child 0, n: int64
          child 1, refusal_rate: double
          child 2, correct_rate: double
      child 1, known_correct_answered: struct<n: int64, refusal_rate: double, correct_rate: double>
          child 0, n: int64
          child 1, refusal_rate: double
          child 2, correct_rate: double
  child 2, shift_minus2: struct<known_refused: st
...
served: double
  child 3, delta: double
  child 4, result: string
provenance: struct<checkpoint: string, base_model: string, adapter: string, direction_file: string, direction_sh (... 78 chars omitted)
  child 0, checkpoint: string
  child 1, base_model: string
  child 2, adapter: string
  child 3, direction_file: string
  child 4, direction_sha256: string
  child 5, runner: string
  child 6, docker_image: string
  child 7, container_run_utc: string
instrument_config_sha256: struct<cell.yaml: string, configs/phase3_current_clean_grpo_v2_caution_perp_residual_intervention.ya (... 204 chars omitted)
  child 0, cell.yaml: string
  child 1, configs/phase3_current_clean_grpo_v2_caution_perp_residual_intervention.yaml: string
  child 2, configs/phase3_current_clean_grpo_v2_caution_residual_intervention.yaml: string
  child 3, configs/phase3_current_clean_grpo_v2_known_overrefusal_native_l26_coeff_sweep.yaml: string
  child 4, gates.yaml: string
license_gate_excluded: struct<>
amendment_path: string
generation_date: string
shape: string
repo_commit_sha: string
sources_present: struct<>
files: struct<config1_caution_residual_intervention_summary.json: string, config2_caution_perp_residual_int (... 77 chars omitted)
  child 0, config1_caution_residual_intervention_summary.json: string
  child 1, config2_caution_perp_residual_intervention_summary.json: string
  child 2, config3_l26_coeff_sweep_metrics.json: string
experiment_slug: string
excluded: list<item: null>
  child 0, item: null
to
{'amendment_path': Value('string'), 'excluded': List(Value('null')), 'experiment_slug': Value('string'), 'files': {'config1_caution_residual_intervention_summary.json': Value('string'), 'config2_caution_perp_residual_intervention_summary.json': Value('string'), 'config3_l26_coeff_sweep_metrics.json': Value('string')}, 'generation_date': Value('string'), 'instrument_config_sha256': {'cell.yaml': Value('string'), 'configs/phase3_current_clean_grpo_v2_caution_perp_residual_intervention.yaml': Value('string'), 'configs/phase3_current_clean_grpo_v2_caution_residual_intervention.yaml': Value('string'), 'configs/phase3_current_clean_grpo_v2_known_overrefusal_native_l26_coeff_sweep.yaml': Value('string'), 'gates.yaml': Value('string')}, 'license_gate_excluded': {}, 'repo_commit_sha': Value('string'), 'shape': Value('string'), 'sources_present': {}}
because column names don't match
Traceback:    Traceback (most recent call last):
                File "/src/services/worker/src/worker/utils.py", line 147, 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 127, 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 483, 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 2840, in __iter__
                  for key, example in ex_iterable:
                                      ^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2373, in __iter__
                  for key, pa_table in self._iter_arrow():
                                       ~~~~~~~~~~~~~~~~^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2398, 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 536, in _iter_arrow
                  for key, pa_table in iterator:
                                       ^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 419, 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 343, 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 2378, in table_cast
                  return cast_table_to_schema(table, schema)
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2306, in cast_table_to_schema
                  raise CastError(
                  ...<3 lines>...
                  )
              datasets.table.CastError: Couldn't cast
              config: string
              config_label: string
              direction_layer: int64
              fingerprint: string
              total_units: int64
              units_generated: int64
              units_resumed: int64
              row_count_per_arm: int64
              arms: list<item: string>
                child 0, item: string
              by_arm: struct<baseline: struct<known_refused: struct<n: int64, refusal_rate: double, correct_rate: double>, (... 635 chars omitted)
                child 0, baseline: struct<known_refused: struct<n: int64, refusal_rate: double, correct_rate: double>, known_correct_an (... 69 chars omitted)
                    child 0, known_refused: struct<n: int64, refusal_rate: double, correct_rate: double>
                        child 0, n: int64
                        child 1, refusal_rate: double
                        child 2, correct_rate: double
                    child 1, known_correct_answered: struct<n: int64, refusal_rate: double, correct_rate: double>
                        child 0, n: int64
                        child 1, refusal_rate: double
                        child 2, correct_rate: double
                child 1, ablate: struct<known_refused: struct<n: int64, refusal_rate: double, correct_rate: double>, known_correct_an (... 69 chars omitted)
                    child 0, known_refused: struct<n: int64, refusal_rate: double, correct_rate: double>
                        child 0, n: int64
                        child 1, refusal_rate: double
                        child 2, correct_rate: double
                    child 1, known_correct_answered: struct<n: int64, refusal_rate: double, correct_rate: double>
                        child 0, n: int64
                        child 1, refusal_rate: double
                        child 2, correct_rate: double
                child 2, shift_minus2: struct<known_refused: st
              ...
              served: double
                child 3, delta: double
                child 4, result: string
              provenance: struct<checkpoint: string, base_model: string, adapter: string, direction_file: string, direction_sh (... 78 chars omitted)
                child 0, checkpoint: string
                child 1, base_model: string
                child 2, adapter: string
                child 3, direction_file: string
                child 4, direction_sha256: string
                child 5, runner: string
                child 6, docker_image: string
                child 7, container_run_utc: string
              instrument_config_sha256: struct<cell.yaml: string, configs/phase3_current_clean_grpo_v2_caution_perp_residual_intervention.ya (... 204 chars omitted)
                child 0, cell.yaml: string
                child 1, configs/phase3_current_clean_grpo_v2_caution_perp_residual_intervention.yaml: string
                child 2, configs/phase3_current_clean_grpo_v2_caution_residual_intervention.yaml: string
                child 3, configs/phase3_current_clean_grpo_v2_known_overrefusal_native_l26_coeff_sweep.yaml: string
                child 4, gates.yaml: string
              license_gate_excluded: struct<>
              amendment_path: string
              generation_date: string
              shape: string
              repo_commit_sha: string
              sources_present: struct<>
              files: struct<config1_caution_residual_intervention_summary.json: string, config2_caution_perp_residual_int (... 77 chars omitted)
                child 0, config1_caution_residual_intervention_summary.json: string
                child 1, config2_caution_perp_residual_intervention_summary.json: string
                child 2, config3_l26_coeff_sweep_metrics.json: string
              experiment_slug: string
              excluded: list<item: null>
                child 0, item: null
              to
              {'amendment_path': Value('string'), 'excluded': List(Value('null')), 'experiment_slug': Value('string'), 'files': {'config1_caution_residual_intervention_summary.json': Value('string'), 'config2_caution_perp_residual_intervention_summary.json': Value('string'), 'config3_l26_coeff_sweep_metrics.json': Value('string')}, 'generation_date': Value('string'), 'instrument_config_sha256': {'cell.yaml': Value('string'), 'configs/phase3_current_clean_grpo_v2_caution_perp_residual_intervention.yaml': Value('string'), 'configs/phase3_current_clean_grpo_v2_caution_residual_intervention.yaml': Value('string'), 'configs/phase3_current_clean_grpo_v2_known_overrefusal_native_l26_coeff_sweep.yaml': Value('string'), 'gates.yaml': Value('string')}, 'license_gate_excluded': {}, 'repo_commit_sha': Value('string'), 'shape': Value('string'), 'sources_present': {}}
              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.

caution-ablation-rederivation -- aggregate exhaust

Aggregate-only: every file committed under this experiment's analysis-committed/ tree (dose-response tables, direction fits, gate AUROCs, manifests, and any other analysis artifact), copied byte-for-byte. No source question text, aliases, or per-row generation text -- analysis-committed/ never carries those.

HF repo: professorsynapse/eh-caution-ablation-rederivation

Provenance

  • Experiment: experiments/caution-ablation-rederivation
  • Amendment: experiments/caution-ablation-rederivation/AMENDMENT.md
  • Repo commit SHA: 092e425ffbad51022dda36386ff515b5c26f3dde
  • Generated: 2026-08-31T11:35:28.096770+00:00
  • Pinned instrument config sha256:
    • cell.yaml: 15a6a6f582b9c25db657ddb5ec673a605d09b257445bf65a959e0ce9ca23b6af
    • configs/phase3_current_clean_grpo_v2_caution_perp_residual_intervention.yaml: 7e0512432fb9dd36761f5f8bb6f325391aa775027d95be50b963938e8582ff33
    • configs/phase3_current_clean_grpo_v2_caution_residual_intervention.yaml: 9952c6948745f6b30f483bcc6a436d550ca05c7a1453c585fce67da4bba0fb29
    • configs/phase3_current_clean_grpo_v2_known_overrefusal_native_l26_coeff_sweep.yaml: 880798a0a3cd886d4ee70dcf1786ca6c3fbcea049bacfe21fb61bc36516e1d69
    • gates.yaml: cf1fc9f4f80f4ea3284e09d5b49b08a1f07bd50ca673fd021c7ae1ad4470d612

File inventory

3 file(s) total, by top-level path:

  • config1_caution_residual_intervention_summary.json: 1 file(s)
  • config2_caution_perp_residual_intervention_summary.json: 1 file(s)
  • config3_l26_coeff_sweep_metrics.json: 1 file(s)

License-gate exclusions

(none)

Hard-exclusion skips

Files whose relative path under analysis-committed/ matched a structural hard-exclusion pattern and were skipped (not copied):

(none)

Files

See PROVENANCE.json for the full machine-readable file list (relative path -> sha256) and the hard-exclusion skip list with reasons.

Release Boundary

Built by .skills/data-exhaust/scripts/build_exhaust_dataset.py, gated by .skills/data-exhaust/reference/license-gates.md, and verified by .skills/data-exhaust/scripts/verify_exhaust.py before any upload. Hard exclusions (OpenMOSS/Cheng IDK, bridge_llama2_7b_chat) are enforced structurally in code, not only by this table.

Downloads last month
22