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
kit_version: string
created_at: timestamp[s]
files: list<item: struct<path: string, sha256: string, bytes: int64>>
  child 0, item: struct<path: string, sha256: string, bytes: int64>
      child 0, path: string
      child 1, sha256: string
      child 2, bytes: int64
root: string
repo_id: string
input: string
id: string
expectations: list<item: struct<kind: string, text: string>>
  child 0, item: struct<kind: string, text: string>
      child 0, kind: string
      child 1, text: string
to
{'id': Value('string'), 'input': Value('string'), 'expectations': List({'kind': Value('string'), 'text': Value('string')})}
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
              kit_version: string
              created_at: timestamp[s]
              files: list<item: struct<path: string, sha256: string, bytes: int64>>
                child 0, item: struct<path: string, sha256: string, bytes: int64>
                    child 0, path: string
                    child 1, sha256: string
                    child 2, bytes: int64
              root: string
              repo_id: string
              input: string
              id: string
              expectations: list<item: struct<kind: string, text: string>>
                child 0, item: struct<kind: string, text: string>
                    child 0, kind: string
                    child 1, text: string
              to
              {'id': Value('string'), 'input': Value('string'), 'expectations': List({'kind': Value('string'), 'text': 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.

YAML Metadata Warning:empty or missing yaml metadata in repo card

Check out the documentation for more information.

GitGrounded bench

Can GitGrounded tell a breaking prompt or policy change from a harmless one on a real LLM app? Each scenario is one git branch of the triage example. GitGrounded runs it against the baseline with the same locked cases and returns PASS, WARN or FAIL.

  • Breaking changes caught as FAIL: 13/20 (95% CI 43% to 82%)
  • Breaking changes caught as FAIL or WARN: 20/20 (95% CI 84% to 100%)
  • Breaking edits that really changed the app's output on at least one case: 20/20. Of those, caught as FAIL: 13/20 (95% CI 43% to 82%), as FAIL or WARN: 20/20
  • Harmless wording changes wrongly marked FAIL: 0/20 (95% CI 0% to 16%)
  • Harmless wording changes marked FAIL or WARN: 14/20
  • Empty commits marked FAIL or WARN: 0/3
  • Median time per PR check: 54 s
  • Total spend reported by GitGrounded: $1.72
  • App outputs that were not valid JSON objects, base and head together: 1

App model: claude-haiku-4-5. Judge model: claude-haiku-4-5. Locked cases: 20. Baseline commit 73f9c48c1931.

A case counts as changed when its category, priority, policy line or assertion results differ between baseline and head. This check reads the raw transcripts and does not use GitGrounded's verdict.

scenario expected verdict cases with changed output fail cases warn cases seconds
bad-01-drop-citation-rule FAIL WARN 9 0 10 111.7
bad-02-drop-policy-field FAIL WARN 19 0 3 65.3
bad-03-remove-grounding FAIL WARN 12 0 7 67.7
bad-04-markdown-output FAIL FAIL 9 1 1 96.1
bad-05-rename-category-field FAIL FAIL 20 20 0 104.5
bad-06-priority-enum FAIL FAIL 20 20 0 113.5
bad-07-guess-when-uncovered FAIL FAIL 11 1 7 62.8
bad-08-allow-invention FAIL WARN 7 0 2 61.9
bad-09-no-policy-context FAIL FAIL 20 19 0 71.4
bad-10-always-low-priority FAIL WARN 16 0 6 64.8
bad-11-answer-in-french FAIL FAIL 9 5 15 68.7
bad-12-promise-full-refund FAIL WARN 5 0 2 57.1
bad-13-one-word-answer FAIL FAIL 11 1 19 63.1
bad-14-paraphrase-policy-line FAIL WARN 16 0 5 62.5
bad-15-truncate-policy FAIL FAIL 15 9 5 67.2
bad-16-refund-window-typo FAIL FAIL 4 2 4 57.9
bad-17-technical-low-priority FAIL FAIL 9 2 3 62.5
bad-18-enterprise-sla-typo FAIL FAIL 3 2 0 56.5
bad-19-deflect-refunds FAIL FAIL 7 4 3 62.1
bad-20-drop-answer-field FAIL FAIL 16 11 3 63.2
safe-01-mid-size PASS WARN 4 0 4 48.0
safe-02-support-assistant PASS WARN 5 0 2 42.4
safe-03-brief-answer PASS WARN 4 0 2 43.3
safe-04-reply-json PASS PASS 2 0 0 41.4
safe-05-make-up PASS WARN 4 0 2 42.3
safe-06-only-source PASS PASS 3 0 0 42.9
safe-07-each-message PASS PASS 2 0 0 41.2
safe-08-precisely PASS PASS 3 0 0 41.0
safe-09-and-no-extra PASS PASS 0 0 0 36.1
safe-10-state-that PASS WARN 3 0 1 19.8
safe-11-policy-is-below PASS WARN 9 0 7 46.9
safe-12-exact-policy-line PASS WARN 3 0 1 44.2
safe-13-that-you-relied PASS WARN 1 0 1 39.0
safe-14-strictly-grounded PASS WARN 4 0 2 42.6
safe-15-be-polite PASS WARN 5 0 2 50.8
safe-16-policy-header PASS WARN 4 0 5 54.0
safe-17-config-comment PASS WARN 1 0 1 39.4
safe-18-for-the-customer PASS WARN 6 0 4 40.7
safe-19-blank-line PASS WARN 4 0 3 46.0
safe-20-never-invent PASS PASS 1 0 0 42.2
null-01-empty-commit PASS PASS 0 0 0 50.1
null-02-empty-commit PASS PASS 0 0 0 51.3
null-03-empty-commit PASS PASS 0 0 0 70.2

Every branch diff is in scenarios.jsonl, and every full GitGrounded result, with each case transcript and judge reason, is in runs/.

Reproduce

git clone https://github.com/amareshhebbar/gitgrounded && cd gitgrounded
pip install -e ".[sign]"
cp bench/proof/.env.example bench/proof/.env
bash bench/proof/run.sh

Verify

python -c "from huggingface_hub import snapshot_download as s; s('AmareshHebbar/GitGrounded_Bench', repo_type='dataset', local_dir='bench/proof/results')"
python bench/proof/verify.py

Manifest root: 38e77fea7e5ffffd5dfd5c95be50f2dcb75937bcf42ee334cad26c810412f495

Downloads last month
41