{
  "url": "https://appliedfrontier.nl/en/research/signal-81f04d25d61707a7/",
  "title": "Reward Granularity in RLVR: Comparing Process and Outcome Reward Structures for Mathematical Reasoning in Small Language Models",
  "published": "2026-07-03",
  "excerpt": "Reinforcement Learning with Verifiable Rewards (RLVR) has emerged as a promising paradigm for improving mathematical reasoning in language models. Yet most RLVR work rewards only",
  "content_sha256": "92b2d6faa7233fd7740f02df9f76cdf70d271971dccf3db5a1fbee835dbe5a3b",
  "extracted_chars": 1563,
  "artifact_links": [
    "https://arxiv.org/abs/2607.02869"
  ],
  "arxiv_id": "2607.02869",
  "id": "signal-81f04d25d61707a7",
  "org": "Research community",
  "topic": "Mathematical reasoning",
  "source_kind": "preprint",
  "status": "unreviewed-source-report",
  "analysis_method": "deterministic_metadata_triage_v1",
  "novelty_verified": false,
  "proof_checked": false,
  "peer_review_verified": false,
  "matched_terms": [
    "mathematical",
    "reasoning"
  ],
  "checked_at": "2026-10-29T22:08:08+00:00",
  "first_seen": "2026-10-29T21:52:00+00:00",
  "source_hash": "bc07ad1e84a50ddff7b8cf5aeed05e48eebe7a45f9db4f7fa536a4ce41ac4d2d",
  "source_url": "https://arxiv.org/abs/2607.02869",
  "language": "en",
  "indexing": "noindex_until_editorially_enriched"
}