{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:UDLFAG2B7A46W3H7LB2RUCD3D4","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"d14dbf0a3b6d1cbe75b6f97fc24d666061829bdaafd3036adf4687a76ca40ddb","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-16T14:55:12Z","title_canon_sha256":"d7a0d58a5af3dcdca97a65fc5a3285d611baf18f0803c7355c6a7bf183dc665c"},"schema_version":"1.0","source":{"id":"2605.17026","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.17026","created_at":"2026-05-20T00:03:36Z"},{"alias_kind":"arxiv_version","alias_value":"2605.17026v1","created_at":"2026-05-20T00:03:36Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.17026","created_at":"2026-05-20T00:03:36Z"},{"alias_kind":"pith_short_12","alias_value":"UDLFAG2B7A46","created_at":"2026-05-20T00:03:36Z"},{"alias_kind":"pith_short_16","alias_value":"UDLFAG2B7A46W3H7","created_at":"2026-05-20T00:03:36Z"},{"alias_kind":"pith_short_8","alias_value":"UDLFAG2B","created_at":"2026-05-20T00:03:36Z"}],"graph_snapshots":[{"event_id":"sha256:82de3fb0f9a15ce3d9750a0671acb37b3c14c1891444183b4a1f92f19e47adb0","target":"graph","created_at":"2026-05-20T00:03:36Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":4,"items":[{"attestation":"unclaimed","claim_id":"C1","kind":"strongest_claim","source":"verdict.strongest_claim","status":"machine_extracted","text":"We hypothesize that this behavior is driven by properties of the fine-tuning data, specifically related to decision points or 'forks in the road' scenarios where model faces indecipherable patterns with multiple valid reasoning paths."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"The controlled case studies using graph branching and reasoning modes accurately capture the decision-point dynamics present in real fine-tuning datasets for reasoning models."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"Coverage shrinkage after SFT in reasoning models correlates with prevalence of decision-point scenarios in data and can be partially mitigated by targeted data synthesis and diversity-aware decoding."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"Fine-tuning data with ambiguous decision points causes reasoning models to lose coverage."}],"snapshot_sha256":"41483c2b9196a4088ad8efd4f8907e158f1fd4815a3d7ef31f64c7073f1297c7"},"formal_canon":{"evidence_count":2,"snapshot_sha256":"3f685701b55407544a51549971253411cb6fca79692b4747d6b14b187d8ed0fa"},"integrity":{"available":true,"clean":true,"detectors_run":[{"findings_count":0,"name":"doi_title_agreement","ran_at":"2026-05-19T21:01:19.007508Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"doi_compliance","ran_at":"2026-05-19T20:40:42.638744Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"citation_quote_validity","ran_at":"2026-05-19T19:49:44.685131Z","status":"skipped","version":"0.1.0"},{"findings_count":0,"name":"cited_work_retraction","ran_at":"2026-05-19T18:51:57.643361Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"claim_evidence","ran_at":"2026-05-19T18:41:56.177593Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"ai_meta_artifact","ran_at":"2026-05-19T18:33:23.494370Z","status":"skipped","version":"1.0.0"}],"endpoint":"/pith/2605.17026/integrity.json","findings":[],"snapshot_sha256":"27b76f4a138faead9813f5402b880aaad5552ffd2df01777d3b1b69f902826e1","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Recent progress in large language models has led to the emergence of reasoning models, which have shown strong performance on complex tasks through specialized fine-tuning procedures. While these methods reliably improve pass@1 accuracy, prior works have observed that they show a coverage shrinkage behavior, where pass@k degrades relative to the base model. In this paper, we investigate the reasoning shrinkage arise under SFT-based post-training. We hypothesize that this behavior is driven by properties of the fine-tuning data, specifically related to decision points or \"forks in the road\" sce","authors_text":"Chandan K Reddy, Khoa D Doan, Nan Zhang, Ngoc-Hieu Nguyen, Parshin Shojaee, Phuc Minh Nguyen, Rui Zhang","cross_cats":[],"headline":"Fine-tuning data with ambiguous decision points causes reasoning models to lose coverage.","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-16T14:55:12Z","title":"Why Do Reasoning Models Lose Coverage? The Role of Data and Forks in the Road"},"references":{"count":19,"internal_anchors":1,"resolved_work":19,"sample":[{"cited_arxiv_id":"2110.14168","doi":"10.1038/s41586-025-09422-z","is_internal_anchor":true,"ref_index":1,"title":"Training Verifiers to Solve Math Word Problems","work_id":"acab1aa8-b4d6-40e0-a3ee-25341701dca2","year":2021},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":2,"title":"Substitutep=l+7 into the target expression, yieldings=l+18","work_id":"2e758a34-c102-4f90-b378-c27e479f1674","year":null},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":3,"title":"Substitutel=m+5 into the target expression, yieldings=m+23","work_id":"ab49658f-2a2f-4199-a457-12ca2d68223d","year":null},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":4,"title":"Substitutem=f+9 into the target expression, yieldings=f+32","work_id":"536dd746-212d-446f-99d1-bd89c04145cc","year":null},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":5,"title":"Substitutef=g+11 into the target expression, yieldings=g+43","work_id":"0164f060-f692-44ca-a74f-c3eadd23e986","year":null}],"snapshot_sha256":"19342224d08837088c7bbe151169a74f4f89bce03a87e0ace330765a72fd7108"},"source":{"id":"2605.17026","kind":"arxiv","version":1},"verdict":{"created_at":"2026-05-19T20:28:40.256953Z","id":"4e675cf6-3703-446b-a1ff-9b7bc3112c4e","model_set":{"reader":"grok-4.3"},"one_line_summary":"Coverage shrinkage after SFT in reasoning models correlates with prevalence of decision-point scenarios in data and can be partially mitigated by targeted data synthesis and diversity-aware decoding.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"Fine-tuning data with ambiguous decision points causes reasoning models to lose coverage.","strongest_claim":"We hypothesize that this behavior is driven by properties of the fine-tuning data, specifically related to decision points or 'forks in the road' scenarios where model faces indecipherable patterns with multiple valid reasoning paths.","weakest_assumption":"The controlled case studies using graph branching and reasoning modes accurately capture the decision-point dynamics present in real fine-tuning datasets for reasoning models."}},"verdict_id":"4e675cf6-3703-446b-a1ff-9b7bc3112c4e"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:4220a5b5e383e78b01cc6bdd026acc6b448f0239d849428c1f7e94e103316bb6","target":"record","created_at":"2026-05-20T00:03:36Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"d14dbf0a3b6d1cbe75b6f97fc24d666061829bdaafd3036adf4687a76ca40ddb","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-16T14:55:12Z","title_canon_sha256":"d7a0d58a5af3dcdca97a65fc5a3285d611baf18f0803c7355c6a7bf183dc665c"},"schema_version":"1.0","source":{"id":"2605.17026","kind":"arxiv","version":1}},"canonical_sha256":"a0d6501b41f839eb6cff58751a087b1f00e1514d87ce92f330b0564b43cfde37","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"a0d6501b41f839eb6cff58751a087b1f00e1514d87ce92f330b0564b43cfde37","first_computed_at":"2026-05-20T00:03:36.598466Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-20T00:03:36.598466Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"tgzPH6hd7eVC53evFp6vbTy1jPlyouVENFd3IXPMGYLnu6rFA113vcsES34/hK4ff5AyySIRL+4Me6yMAeR7DQ==","signature_status":"signed_v1","signed_at":"2026-05-20T00:03:36.599256Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.17026","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:4220a5b5e383e78b01cc6bdd026acc6b448f0239d849428c1f7e94e103316bb6","sha256:82de3fb0f9a15ce3d9750a0671acb37b3c14c1891444183b4a1f92f19e47adb0"],"state_sha256":"53ca4de0ec2726fa548d19904426c03aaf5e63167a087060d174e73622bc5c87"}