{
  "$schema": "https://themachinepress.com/schemas/story-v1.schema.json",
  "schema_version": "1.0.0",
  "document_type": "machine_press_story",
  "story": {
    "story_id": "mp-2026-11-03-010",
    "source_story_id": "tmp-story-rlvr-strategy-collapse",
    "edition_id": "mp-2026-11-03-morning-0116",
    "edition_url": "https://themachinepress.com/edition/2026-11-03",
    "position": 10,
    "story_type": "dispatch",
    "section": "research",
    "editorial_classification": "editorial",
    "headline": "The Reasoning Trace Works—Until the Problem Gets Harder",
    "slug": "the-reasoning-trace-works-until-the-problem-gets-harder",
    "dek": "An RLVR model solved in-distribution equations with structured derivations, then abandoned that strategy for brute-force enumeration under harder shifts.",
    "summary": "An RLVR model solved in-distribution equations with structured derivations, then abandoned that strategy for brute-force enumeration under harder shifts.",
    "body_text": "A controlled equation-solving setup makes strategy visible: a model must find multiple inputs for a non-bijective string function while tests vary the number of answers and output length. Reinforcement learning with verifiable rewards produced high in-distribution accuracy and orderly derivations, yet under harder out-of-distribution shifts the same model dropped the derivation and resorted to ineffective enumeration. The failure was neither a weak base model nor ordinary reward hacking. Co-training on bijective inversion tasks—different and not uniformly harder—made derivation more certain and restored much of the strategy and accuracy under shift. The finding is a case study, but it warns that outcome success can conceal brittle reasoning policies.",
    "why_it_matters": "An RLVR model solved in-distribution equations with structured derivations, then abandoned that strategy for brute-force enumeration under harder shifts.",
    "limitations": [
      "Co-training on bijective inversion tasks—different and not uniformly harder—made derivation more certain and restored much of the strategy and accuracy under shift."
    ],
    "importance": 8,
    "canonical_url": "https://themachinepress.com/story/mp-2026-11-03-010/the-reasoning-trace-works-until-the-problem-gets-harder",
    "json_url": "https://themachinepress.com/story/mp-2026-11-03-010.json",
    "first_published_at": "2026-11-03T09:00:00.000-05:00",
    "modified_at": "2026-11-03T09:00:00.000-05:00",
    "content_status": "new",
    "is_carryover": false,
    "carryover_reason": null,
    "key_claims": [
      {
        "claim_id": "claim-mp-2026-11-03-010-001",
        "text": "An RLVR model solved in-distribution equations with structured derivations, then abandoned that strategy for brute-force enumeration under harder shifts.",
        "source_ids": [
          "source-2026-11-03-010"
        ],
        "qualification": "Co-training on bijective inversion tasks—different and not uniformly harder—made derivation more certain and restored much of the strategy and accuracy under shift."
      }
    ],
    "source_ids": [
      "source-2026-11-03-010"
    ],
    "tags": [
      "RLVR",
      "reasoning",
      "generalization",
      "evaluation"
    ],
    "image_url": null,
    "corrections": []
  },
  "sources": [
    {
      "source_id": "source-2026-11-03-010",
      "title": "arXiv preprint 2611.03627, original v1",
      "publisher": "arXiv",
      "url": "https://arxiv.org/abs/2611.03627v1",
      "canonical_url": "https://arxiv.org/abs/2611.03627v1",
      "source_type": "primary_research",
      "is_primary_source": true,
      "published_at": "2026-11-02T13:53:54.000-05:00",
      "accessed_at": "2026-11-03T08:18:17.000-05:00",
      "supports_claim_ids": [
        "claim-mp-2026-11-03-010-001"
      ]
    }
  ],
  "corrections": [],
  "publisher": {
    "name": "The Machine Press",
    "url": "https://themachinepress.com",
    "description": "A daily newspaper for the age of artificial intelligence."
  },
  "cite_this_report": {
    "title": "The Reasoning Trace Works—Until the Problem Gets Harder",
    "publisher": "The Machine Press",
    "published_at": "2026-11-03T09:00:00.000-05:00",
    "canonical_url": "https://themachinepress.com/story/mp-2026-11-03-010/the-reasoning-trace-works-until-the-problem-gets-harder"
  }
}