Datasets:

Rajveer-code
/

flipbudget-results

Error code:   StreamingRowsError
Exception:    ValueError
Message:      Expected object or value
Traceback:    Traceback (most recent call last):
                File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 290, in _generate_tables
                  pa_table = paj.read_json(
                      io.BytesIO(batch), read_options=paj.ReadOptions(block_size=block_size)
                  )
                File "pyarrow/_json.pyx", line 342, in pyarrow._json.read_json
                File "pyarrow/error.pxi", line 155, in pyarrow.lib.pyarrow_internal_check_status
                File "pyarrow/error.pxi", line 92, in pyarrow.lib.check_status
                  raise convert_status(status)
              pyarrow.lib.ArrowInvalid: JSON parse error: Column(/findings/[]/what_changed) changed from string to array in row 0
              
              During handling of the above exception, another exception occurred:
              
              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 304, in _generate_tables
                  batch = json_encode_fields_in_json_lines(original_batch, json_field_paths)
                File "/usr/local/lib/python3.14/site-packages/datasets/utils/json.py", line 111, in json_encode_fields_in_json_lines
                  examples = [ujson_loads(line) for line in original_batch.splitlines()]
                              ~~~~~~~~~~~^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/utils/json.py", line 20, in ujson_loads
                  return pd.io.json.ujson_loads(*args, **kwargs)
                         ~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^
              ValueError: Expected object or value

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.

flipbudget-results

Real, verified results from a scorer-induced partial-identification analysis of benchmark comparisons. Companion dataset to the code and theory at github.com/Rajveer-code/flipbudget.

What this is

Every file here is a script output committed alongside the script that produced it — no number in this dataset is asserted without a runnable, version-controlled script behind it. See the GitHub repo's commit history for the full derivation, including one real bug found and fixed in a coverage simulation (a wrong ground-truth reference), and one design-weighted accuracy estimator that its own pre-declared sanity check correctly rejected before a corrected version replaced it.

Files

results/flipbudget/ — this program's own derived findings:

e1_pilot_case_a.json — Case A (non-differential) attenuation on all 378 real MATH-Hard pairs
e1_case_b_TRUE.json — the real GO/NO-GO result: per-model true accuracy (from raw generations, not a reweighted sample), and per-pair flip budgets
e4_mathhard_per_model.json — real per-model false-credit/false-miss rates (27 of 28 roster models), partial-pooled
e5_leaderboard.json — which of the 378 pairs remain identified under the real, currently-measured differential error
e6_version_drift_summary.json — git-history evidence that the scoring pipeline itself drifts version to version, in the same dimension this program studies
d4_data_sources.json — audit of which benchmarks had usable data before any pilot was scoped

results/analysis/ — the upstream snapshot these scripts consume (model responses, extractor outputs, human-audited margins) from the companion project auditing answer-extraction failure in LLM benchmark harnesses.

Headline finding

On the real MATH-Hard roster (28 models, 378 pairwise comparisons): 16.1% of pairs have a flip budget below the differential misclassification already observed across the roster. 99.4% remain identified under today's actual measured error — the two numbers are not in tension; they answer different questions (see the GitHub repo's E1/E5 commit messages).

License

Apache-2.0, matching the code repository.

Downloads last month: 47

Total file size:

5.97 MB