Hugging Face's logo

Datasets:
juspay
/
xor-26b-a4b-v1.2-decision-index

Tags:
decision-index
License:
Dataset card Files
xet
 Community
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
served_model_dir: string
release: string
files: struct<chat_template.jinja: struct<served: string, release: string, identical: bool>, config.json: s (... 646 chars omitted)
  child 0, chat_template.jinja: struct<served: string, release: string, identical: bool>
      child 0, served: string
      child 1, release: string
      child 2, identical: bool
  child 1, config.json: struct<served: string, release: string, identical: bool>
      child 0, served: string
      child 1, release: string
      child 2, identical: bool
  child 2, generation_config.json: struct<served: string, release: string, identical: bool>
      child 0, served: string
      child 1, release: string
      child 2, identical: bool
  child 3, model-00001-of-00002.safetensors: struct<served: string, release: string, identical: bool>
      child 0, served: string
      child 1, release: string
      child 2, identical: bool
  child 4, model-00002-of-00002.safetensors: struct<served: string, release: string, identical: bool>
      child 0, served: string
      child 1, release: string
      child 2, identical: bool
  child 5, model.safetensors.index.json: struct<served: string, release: string, identical: bool>
      child 0, served: string
      child 1, release: string
      child 2, identical: bool
  child 6, processor_config.json: struct<served: string, release: string, identical: bool>
      child 0, served: string
      child 1, release: string
      child 2, identical: bool
  child 7, tokenizer.json: s
...
iSarcasmEval-A-Ar: struct<metric: string, score: null, scored_requests: int64>
              child 0, metric: string
              child 1, score: null
              child 2, scored_requests: int64
          child 3, iSarcasmEval-A-En: struct<metric: string, score: null, scored_requests: int64>
              child 0, metric: string
              child 1, score: null
              child 2, scored_requests: int64
          child 4, iSarcasmEval-B-En: struct<metric: string, score: null, scored_requests: int64>
              child 0, metric: string
              child 1, score: null
              child 2, scored_requests: int64
          child 5, iSarcasmEval-C-Ar: struct<metric: string, score: null, scored_requests: int64>
              child 0, metric: string
              child 1, score: null
              child 2, scored_requests: int64
          child 6, iSarcasmEval-C-En: struct<metric: string, score: null, scored_requests: int64>
              child 0, metric: string
              child 1, score: null
              child 2, scored_requests: int64
          child 7, GSM8K-10choice: struct<metric: string, score: double, scored_requests: int64>
              child 0, metric: string
              child 1, score: double
              child 2, scored_requests: int64
          child 8, GSM8K-4choice: struct<metric: string, score: double, scored_requests: int64>
              child 0, metric: string
              child 1, score: double
              child 2, scored_requests: int64
to
{'engine': Value('string'), 'counts': {'ok': Value('int64')}, 'successful_request_latency_ms': {'median': Value('float64'), 'p95': Value('float64'), 'mean': Value('float64')}, 'benchmarks': List({'catalog_id': Value('int64'), 'dataset': Value('string'), 'requests': Value('int64'), 'answered': Value('int64'), 'unsupported': Value('int64'), 'errors': Value('int64'), 'abstained': Value('int64'), 'pending': Value('int64'), 'scored_requests': Value('int64'), 'metric': Value('string'), 'score': Value('float64'), 'reference_same_cases': Value('null'), 'median_ms': Value('float64'), 'detail': {'field_accuracy': Value('float64'), 'case_exact_accuracy': Value('float64'), 'custom_metrics': {'ndcg_at_10': Value('float64'), 'mrr': Value('float64'), 'recall_at_10': Value('float64'), 'candidate_recall': Value('float64'), 'scorable_candidate_recall': Value('float64'), 'candidates_scored': Value('float64'), 'candidates_retrieved': Value('int64'), 'bm25_ndcg_at_10': Value('float64'), 'quality_quality': Value('float64'), 'quality_cost_usd': Value('float64'), 'quality_utility': Value('float64'), 'quality_oracle_optimal': Value('float64'), 'quality_utility_regret': Value('float64'), 'cost_aware_quality': Value('float64'), 'cost_aware_cost_usd': Value('float64'), 'cost_aware_utility': Value('float64'), 'cost_aware_oracle_optimal': Value('float64'), 'cost_aware_utility_regret': Value('float64'), 'value_regret': Value('float64'), 'normalized_rank_regret': Value('float64')}, 'positive_micro_f1': Valu
...
Value('int64'), 'mean': Value('int64')}}, 'cluster_macro_accuracy': Value('float64'), 'positive_f1_by_field': {'sarcastic': Value('float64'), 'sarcasm': Value('float64'), 'irony': Value('float64'), 'satire': Value('float64'), 'understatement': Value('float64'), 'overstatement': Value('float64'), 'rhetorical_question': Value('float64')}, 'category_macro_f1': Value('float64'), 'scored_fields': Value('int64'), 'chance_on_rows': Value('float64')}, 'tracks': {'RouterBench-0shot': {'metric': Value('string'), 'score': Value('float64'), 'scored_requests': Value('int64')}, 'RouterBench-5shot': {'metric': Value('string'), 'score': Value('float64'), 'scored_requests': Value('int64')}, 'iSarcasmEval-A-Ar': {'metric': Value('string'), 'score': Value('null'), 'scored_requests': Value('int64')}, 'iSarcasmEval-A-En': {'metric': Value('string'), 'score': Value('null'), 'scored_requests': Value('int64')}, 'iSarcasmEval-B-En': {'metric': Value('string'), 'score': Value('null'), 'scored_requests': Value('int64')}, 'iSarcasmEval-C-Ar': {'metric': Value('string'), 'score': Value('null'), 'scored_requests': Value('int64')}, 'iSarcasmEval-C-En': {'metric': Value('string'), 'score': Value('null'), 'scored_requests': Value('int64')}, 'GSM8K-10choice': {'metric': Value('string'), 'score': Value('float64'), 'scored_requests': Value('int64')}, 'GSM8K-4choice': {'metric': Value('string'), 'score': Value('float64'), 'scored_requests': Value('int64')}}}), 'note': Value('string'), 'edition': Value('string')}
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 2951, in __iter__
                  for key, example in ex_iterable:
                                      ^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2461, in __iter__
                  for key, pa_table in self._iter_arrow():
                                       ~~~~~~~~~~~~~~~~^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2486, 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 547, in _iter_arrow
                  for key, pa_table in iterator:
                                       ^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 430, 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
              served_model_dir: string
              release: string
              files: struct<chat_template.jinja: struct<served: string, release: string, identical: bool>, config.json: s (... 646 chars omitted)
                child 0, chat_template.jinja: struct<served: string, release: string, identical: bool>
                    child 0, served: string
                    child 1, release: string
                    child 2, identical: bool
                child 1, config.json: struct<served: string, release: string, identical: bool>
                    child 0, served: string
                    child 1, release: string
                    child 2, identical: bool
                child 2, generation_config.json: struct<served: string, release: string, identical: bool>
                    child 0, served: string
                    child 1, release: string
                    child 2, identical: bool
                child 3, model-00001-of-00002.safetensors: struct<served: string, release: string, identical: bool>
                    child 0, served: string
                    child 1, release: string
                    child 2, identical: bool
                child 4, model-00002-of-00002.safetensors: struct<served: string, release: string, identical: bool>
                    child 0, served: string
                    child 1, release: string
                    child 2, identical: bool
                child 5, model.safetensors.index.json: struct<served: string, release: string, identical: bool>
                    child 0, served: string
                    child 1, release: string
                    child 2, identical: bool
                child 6, processor_config.json: struct<served: string, release: string, identical: bool>
                    child 0, served: string
                    child 1, release: string
                    child 2, identical: bool
                child 7, tokenizer.json: s
              ...
              iSarcasmEval-A-Ar: struct<metric: string, score: null, scored_requests: int64>
                            child 0, metric: string
                            child 1, score: null
                            child 2, scored_requests: int64
                        child 3, iSarcasmEval-A-En: struct<metric: string, score: null, scored_requests: int64>
                            child 0, metric: string
                            child 1, score: null
                            child 2, scored_requests: int64
                        child 4, iSarcasmEval-B-En: struct<metric: string, score: null, scored_requests: int64>
                            child 0, metric: string
                            child 1, score: null
                            child 2, scored_requests: int64
                        child 5, iSarcasmEval-C-Ar: struct<metric: string, score: null, scored_requests: int64>
                            child 0, metric: string
                            child 1, score: null
                            child 2, scored_requests: int64
                        child 6, iSarcasmEval-C-En: struct<metric: string, score: null, scored_requests: int64>
                            child 0, metric: string
                            child 1, score: null
                            child 2, scored_requests: int64
                        child 7, GSM8K-10choice: struct<metric: string, score: double, scored_requests: int64>
                            child 0, metric: string
                            child 1, score: double
                            child 2, scored_requests: int64
                        child 8, GSM8K-4choice: struct<metric: string, score: double, scored_requests: int64>
                            child 0, metric: string
                            child 1, score: double
                            child 2, scored_requests: int64
              to
              {'engine': Value('string'), 'counts': {'ok': Value('int64')}, 'successful_request_latency_ms': {'median': Value('float64'), 'p95': Value('float64'), 'mean': Value('float64')}, 'benchmarks': List({'catalog_id': Value('int64'), 'dataset': Value('string'), 'requests': Value('int64'), 'answered': Value('int64'), 'unsupported': Value('int64'), 'errors': Value('int64'), 'abstained': Value('int64'), 'pending': Value('int64'), 'scored_requests': Value('int64'), 'metric': Value('string'), 'score': Value('float64'), 'reference_same_cases': Value('null'), 'median_ms': Value('float64'), 'detail': {'field_accuracy': Value('float64'), 'case_exact_accuracy': Value('float64'), 'custom_metrics': {'ndcg_at_10': Value('float64'), 'mrr': Value('float64'), 'recall_at_10': Value('float64'), 'candidate_recall': Value('float64'), 'scorable_candidate_recall': Value('float64'), 'candidates_scored': Value('float64'), 'candidates_retrieved': Value('int64'), 'bm25_ndcg_at_10': Value('float64'), 'quality_quality': Value('float64'), 'quality_cost_usd': Value('float64'), 'quality_utility': Value('float64'), 'quality_oracle_optimal': Value('float64'), 'quality_utility_regret': Value('float64'), 'cost_aware_quality': Value('float64'), 'cost_aware_cost_usd': Value('float64'), 'cost_aware_utility': Value('float64'), 'cost_aware_oracle_optimal': Value('float64'), 'cost_aware_utility_regret': Value('float64'), 'value_regret': Value('float64'), 'normalized_rank_regret': Value('float64')}, 'positive_micro_f1': Valu
              ...
              Value('int64'), 'mean': Value('int64')}}, 'cluster_macro_accuracy': Value('float64'), 'positive_f1_by_field': {'sarcastic': Value('float64'), 'sarcasm': Value('float64'), 'irony': Value('float64'), 'satire': Value('float64'), 'understatement': Value('float64'), 'overstatement': Value('float64'), 'rhetorical_question': Value('float64')}, 'category_macro_f1': Value('float64'), 'scored_fields': Value('int64'), 'chance_on_rows': Value('float64')}, 'tracks': {'RouterBench-0shot': {'metric': Value('string'), 'score': Value('float64'), 'scored_requests': Value('int64')}, 'RouterBench-5shot': {'metric': Value('string'), 'score': Value('float64'), 'scored_requests': Value('int64')}, 'iSarcasmEval-A-Ar': {'metric': Value('string'), 'score': Value('null'), 'scored_requests': Value('int64')}, 'iSarcasmEval-A-En': {'metric': Value('string'), 'score': Value('null'), 'scored_requests': Value('int64')}, 'iSarcasmEval-B-En': {'metric': Value('string'), 'score': Value('null'), 'scored_requests': Value('int64')}, 'iSarcasmEval-C-Ar': {'metric': Value('string'), 'score': Value('null'), 'scored_requests': Value('int64')}, 'iSarcasmEval-C-En': {'metric': Value('string'), 'score': Value('null'), 'scored_requests': Value('int64')}, 'GSM8K-10choice': {'metric': Value('string'), 'score': Value('float64'), 'scored_requests': Value('int64')}, 'GSM8K-4choice': {'metric': Value('string'), 'score': Value('float64'), 'scored_requests': Value('int64')}}}), 'note': Value('string'), 'edition': 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.

Xor 26B-A4B v1.2: Decision Index 0.3

A complete run of Xor 26B-A4B v1.2 (juspay/Xor-26B-A4B, tag v1.2 @ e4d0e994, BF16) on the public Decision Index 0.3 suite.

Decision Index Raw index Breadth skill Hardware
59.46 69.41 58.76 2 x RTX PRO 6000 Blackwell, two servers

The run is complete ("complete": true, 140,178 of 140,178 scoreable requests). All 140,620 requests (including the 442 scoring exclusions) returned ok: no unsupported, error or abstained rows. Nothing was truncated and no options were removed. It is one fresh run from an empty results directory; no rows were resumed or reused from earlier runs or editions.

Area Skill Raw
Knowledge & Reasoning 45.50 57.49
Language Understanding 63.22 73.72
Retrieval & Classification 64.84 73.88
Tools & Automation 72.17 75.72
Arts & Human Taste 51.86 68.59

What was scored

  • Weights. The released juspay/Xor-26B-A4B v1.2 files, downloaded and hash-checked before serving. All 9 weight, index, config and tokenizer files served during the run are byte-identical to the release (harness/release-identity.json).
  • Serving. The release's own bundle, serving/xor-v1.2-serving.tar.gz, started with its run.sh (one Docker Compose project per GPU): SGLang image prakhar1611/xor-sglang@sha256:94c48d2a6cc98dc456cf93f723707ea7dd81dddfe1061e823b348d68bbe8158f (built on SGLang 408d2334), the bundle's SGLang flags (TP1, --context-length 131072 --max-prefill-tokens 131072 --chunked-prefill-size 16384 --mem-fraction-static 0.80), and the bundle's compatibility server xn_base_server.py (sha256 2eed5693…) with two-order scoring and temperatures choice 1.95, noul 1.0, score 1.0.

Files (runs/xor-26b-a4b-v1.2/)

  • results.jsonl.gz: one row per request in the kit format. Each row payload (the suite state and questions) is left out so the suite is not republished; every row keeps its payload_sha256. Re-scoring this file against the 0.3 suite reproduces scores.json, index.json and benchmark-summary.json exactly (harness/strip_and_verify.sh).
  • scores.json, index.json, benchmark-summary.json, score.log: output of decision_index score --edition 0.3.
  • harness/:
    • make_shards.py, shards.json: the 0.3 run list (suite.row_paths filtered by suite.in_edition) split round-robin into 12 shards, with each shard file's sha256 and row count.
    • run_all.sh: one decision_index run --rows shard_NN.jsonl.gz process per shard; even shards against the GPU 4 server, odd shards against the GPU 5 server. The 12 result files were concatenated (140,620 unique run ids, no duplicates) and scored once.
    • shards/shard-NN/: each runner's environment.json, status.json, runner.log. Every runner was started once and completed.
    • strip_and_verify.sh: payload removal and re-score check.
    • release-identity.json: served files versus the v1.2 release.
  • Setup: kit f5a6e4f (edition 0.3, suite hashes verified with decision-index suite verify), unmodified http engine and runner. 2 x NVIDIA RTX PRO 6000 Blackwell Server Edition (96 GB), two servers (TP1, one GPU each), 12 runner processes, about 1 hour 50 minutes.
  • Latency (median 176.6 ms, p95 2,268.4 ms, mean 559.5 ms) was measured with 12 requests in flight across the two servers, not single-client.
Downloads last month
74
Total file size:
58.1 MB