Datasets:

opsd-genrm
/

dedup_filtered_HS3

language:
  - en
license: cc-by-4.0
task_categories:
  - text-classification
pretty_name: Dedup-Filtered HelpSteer3 with Rubrics
size_categories:
  - 10K<n<100K
source_datasets:
  - nvidia/HelpSteer3
tags:
  - reward-model
  - preference
  - rubric
  - sdpo

Dedup-Filtered HelpSteer3 (+ rubrics on train)

A cleaned version of nvidia/HelpSteer3 (preference config) with three modifications:

Filter out domain == "multilingual" and overall_preference == 0 (tie) rows.
Within-split dedup by content hash (sha1(context, response1, response2)), keeping the first occurrence. Upstream HS3 has ~35% byte-identical duplicate rows after the filter step.
Cross-split dedup: drop validation rows whose content hash also appears in train. Upstream HS3 has ~918 of ~1,553 filtered-deduped val rows that are byte-identical to a train row, which would otherwise cause train/val leakage.

Both splits carry a rubric column. Train rubrics were generated by claude-opus-4-6-v1 via an external multicall rubric-gen pipeline and then content-hash joined onto each train row. Each rubric is a judging rubric for the pair (sections: Task, Hard requirements, Discriminative criteria, Pitfalls). Train rubrics are guaranteed non-empty (100% coverage). Validation rubrics are always the empty string — val eval does not use the rubric; the column is present only for HF DatasetDict schema parity.

Sizes

Split	Rows	`rubric`
train	19557	non-empty (100% covered)
validation	635	empty string

Schema

train and validation both inherit the upstream HS3 preference schema (context, response1, response2, overall_preference, individual_preference, domain, …). train adds one column:

rubric: str — evaluation rubric for the pair, produced by the rubric-gen pipeline (sections: Task, Hard requirements, Discriminative criteria, Pitfalls). Guaranteed non-empty.

Dedup details

Content hash: sha1(json.dumps([context, response1, response2], sort_keys=True, ensure_ascii=False)).
Keep-first on first occurrence. Upstream HS3 has 13 groups where (context, response1, response2) are identical but overall_preference / individual_preference differ; for those groups, this dataset keeps the first occurrence's labels.

Train/val leak

The raw HS3 preference config has a substantial train/val content overlap: after standard filters, 918 of 1,553 unique val pairs are also in train. This dataset removes those from val, leaving 635 genuinely-unseen pairs.

License

Inherits upstream HelpSteer3's license. Please cite the original dataset:

@misc{wang2025helpsteer3,
  title={HelpSteer 3},
  author={Wang, Zhilin and others},
  year={2025},
  publisher={NVIDIA},
  url={https://huggingface.co/datasets/nvidia/HelpSteer3}
}