asingh15/fineproofs-prm-context-v2-solution

1---2pretty_name: FineProofs PRM Context v2 - Solution3task_categories:4- text-classification5tags:6- process-reward-model7- theorem-proving8- fineproofs9configs:10- config_name: default11  data_files:12  - split: train13    path: train.parquet14  - split: validation15    path: validation.parquet16---17 18# FineProofs PRM Context v2: Solution19 20This arm uses `solution` context from `same_problem` rollouts with packing policy `whole_solutions_no_truncation`.21 22This is one of nine row-matched context variants built from the verified FineProofs rollout collection. Partial-prefix and complete-response targets both use canonical normalized rubric credit derived from clamped points divided by max points. The `correct` column is only a legacy boolean projection at reward >= 0.5; training uses the dense `reward` target.23 24## Contract25 26- Run ID: `fineproofs_all_qwen35_9b_direct2phase_m32_20260730`27- Collection model: `Qwen/Qwen3.5-9B`28- Collection model revision: `c202236235762e1c871ad0ccb60c8ee5ba337b9a`29- Context arm: `solution`30- Context mode: `solution`31- Context scope: `same_problem`32- Correctness labels in context: `no`33- Verification SHA-256: `c5a3b92bd6196d4a33cded2c1a01870d1963d323351d573b369b7f97f3b7f648`34 35The nine arms have identical row keys, labels, rewards, train/validation problem split, and hard endpoint coverage. Held-out problems are removed before splitting and context selection.36 37## Splits38 39| Split | Rows | Problems |40| --- | ---: | ---: |41| train | 53,457 | 2,542 |42| validation | 2,602 | 128 |43 44The local `val.parquet` is published as `validation.parquet`. `dataset_provenance.json`, `verification.json`, and `_SUCCESS.json` contain the live-input provenance, cross-arm checks, and exact published file fingerprints.45