{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:A3IPCBGAJZN5UKQYTS4ROJU67X","short_pith_number":"pith:A3IPCBGA","schema_version":"1.0","canonical_sha256":"06d0f104c04e5bda2a189cb917269efdedb6fe236704b9bd2954908eb7d8a018","source":{"kind":"arxiv","id":"2606.07190","version":1},"attestation_state":"computed","paper":{"title":"From Correctness to Utility: Gain-Based Prefix Evaluation for LLM Reasoning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Guangnan Ye, Yixin Cao, Yuhang Zhou","submitted_at":"2026-06-05T11:56:50Z","abstract_excerpt":"Reasoning prefixes shape the future trajectory of LLM problem solving, yet existing process reward models usually evaluate them through local step correctness. We argue that correctness is a useful but indirect proxy for the effect we ultimately care about: whether a prefix increases the probability of successful completion. We define this effect as prefix gain, the solve-rate improvement induced by conditioning lightweight student model group on a prefix, and use it to train a Prefix Utility Model (PUM) with a simple pairwise ranking objective. PUM learns outcome-grounded prefix utility and c"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2606.07190","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-06-05T11:56:50Z","cross_cats_sorted":[],"title_canon_sha256":"5f3caa2e8cc224e44c8c0ff353208b71295b23571398a3aacf50e0b862d2f1b1","abstract_canon_sha256":"3920c33890a578c96fc6a0b7fd8723de59e319cb8f6ffa1341972639c9d980f8"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-08T01:04:51.711216Z","signature_b64":"+N1isLRsYkL0zxsxIfHa1R5UKBzA69yJB7kYESgbH9E4gQwxvR6AmklUq6NY5yGMQk65C7YRIGbUQS4F0xqfBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"06d0f104c04e5bda2a189cb917269efdedb6fe236704b9bd2954908eb7d8a018","last_reissued_at":"2026-06-08T01:04:51.710290Z","signature_status":"signed_v1","first_computed_at":"2026-06-08T01:04:51.710290Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"From Correctness to Utility: Gain-Based Prefix Evaluation for LLM Reasoning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Guangnan Ye, Yixin Cao, Yuhang Zhou","submitted_at":"2026-06-05T11:56:50Z","abstract_excerpt":"Reasoning prefixes shape the future trajectory of LLM problem solving, yet existing process reward models usually evaluate them through local step correctness. We argue that correctness is a useful but indirect proxy for the effect we ultimately care about: whether a prefix increases the probability of successful completion. We define this effect as prefix gain, the solve-rate improvement induced by conditioning lightweight student model group on a prefix, and use it to train a Prefix Utility Model (PUM) with a simple pairwise ranking objective. PUM learns outcome-grounded prefix utility and c"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.07190","kind":"arxiv","version":1},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2606.07190/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"2606.07190","created_at":"2026-06-08T01:04:51.710432+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.07190v1","created_at":"2026-06-08T01:04:51.710432+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.07190","created_at":"2026-06-08T01:04:51.710432+00:00"},{"alias_kind":"pith_short_12","alias_value":"A3IPCBGAJZN5","created_at":"2026-06-08T01:04:51.710432+00:00"},{"alias_kind":"pith_short_16","alias_value":"A3IPCBGAJZN5UKQY","created_at":"2026-06-08T01:04:51.710432+00:00"},{"alias_kind":"pith_short_8","alias_value":"A3IPCBGA","created_at":"2026-06-08T01:04:51.710432+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/A3IPCBGAJZN5UKQYTS4ROJU67X","json":"https://pith.science/pith/A3IPCBGAJZN5UKQYTS4ROJU67X.json","graph_json":"https://pith.science/api/pith-number/A3IPCBGAJZN5UKQYTS4ROJU67X/graph.json","events_json":"https://pith.science/api/pith-number/A3IPCBGAJZN5UKQYTS4ROJU67X/events.json","paper":"https://pith.science/paper/A3IPCBGA"},"agent_actions":{"view_html":"https://pith.science/pith/A3IPCBGAJZN5UKQYTS4ROJU67X","download_json":"https://pith.science/pith/A3IPCBGAJZN5UKQYTS4ROJU67X.json","view_paper":"https://pith.science/paper/A3IPCBGA","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.07190&json=true","fetch_graph":"https://pith.science/api/pith-number/A3IPCBGAJZN5UKQYTS4ROJU67X/graph.json","fetch_events":"https://pith.science/api/pith-number/A3IPCBGAJZN5UKQYTS4ROJU67X/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/A3IPCBGAJZN5UKQYTS4ROJU67X/action/timestamp_anchor","attest_storage":"https://pith.science/pith/A3IPCBGAJZN5UKQYTS4ROJU67X/action/storage_attestation","attest_author":"https://pith.science/pith/A3IPCBGAJZN5UKQYTS4ROJU67X/action/author_attestation","sign_citation":"https://pith.science/pith/A3IPCBGAJZN5UKQYTS4ROJU67X/action/citation_signature","submit_replication":"https://pith.science/pith/A3IPCBGAJZN5UKQYTS4ROJU67X/action/replication_record"}},"created_at":"2026-06-08T01:04:51.710432+00:00","updated_at":"2026-06-08T01:04:51.710432+00:00"}