{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:SG6WVWVPS76OFAEKG36XENOW4R","short_pith_number":"pith:SG6WVWVP","schema_version":"1.0","canonical_sha256":"91bd6adaaf97fce2808a36fd7235d6e47df14137556681c304608675d87e3a48","source":{"kind":"arxiv","id":"2607.28457","version":1},"attestation_state":"computed","paper":{"title":"SVR: Self-Verifying Refinement via Joint Verdict-Confidence Reinforcement Learning for Adaptive Test-Time Compute","license":"http://creativecommons.org/publicdomain/zero/1.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.AI","authors_text":"Guangrun Wang, Hongyu Chen, Liang Lin","submitted_at":"2026-07-30T16:20:58Z","abstract_excerpt":"Scaling test-time computation can improve language-model reasoning, but uniform budgets waste computation on easy inputs, while verifier-guided refinement relies on external feedback. We introduce Self-Verifying Refinement (SVR), an oracle-free multi-turn reinforcement learning framework that learns to use self-verification as a compute-control policy. At each turn, the model produces a solution together with a discrete correctness verdict and a confidence score; it retains the current answer only when the verdict is Correct and confidence exceeds a threshold, and otherwise continues refinemen"},"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":"2607.28457","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/publicdomain/zero/1.0/","primary_cat":"cs.AI","submitted_at":"2026-07-30T16:20:58Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"9fb6026db3601a2fa2303d0919ebeae575bb1213902b3ab8163d795b54ede601","abstract_canon_sha256":"1bfdcdc1fbc5fafa188cac9ee39732094526717de94716d801a5dc96b1e97f3d"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"91bd6adaaf97fce2808a36fd7235d6e47df14137556681c304608675d87e3a48","last_reissued_at":"2026-07-31T01:37:40.754737Z","signature_status":"unsigned_v0","first_computed_at":"2026-07-31T01:37:40.754737Z"},"graph_snapshot":{"paper":{"title":"SVR: Self-Verifying Refinement via Joint Verdict-Confidence Reinforcement Learning for Adaptive Test-Time Compute","license":"http://creativecommons.org/publicdomain/zero/1.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.AI","authors_text":"Guangrun Wang, Hongyu Chen, Liang Lin","submitted_at":"2026-07-30T16:20:58Z","abstract_excerpt":"Scaling test-time computation can improve language-model reasoning, but uniform budgets waste computation on easy inputs, while verifier-guided refinement relies on external feedback. We introduce Self-Verifying Refinement (SVR), an oracle-free multi-turn reinforcement learning framework that learns to use self-verification as a compute-control policy. At each turn, the model produces a solution together with a discrete correctness verdict and a confidence score; it retains the current answer only when the verdict is Correct and confidence exceeds a threshold, and otherwise continues refinemen"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.28457","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/2607.28457/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":"2607.28457","created_at":"2026-07-31T01:37:40.757977+00:00"},{"alias_kind":"arxiv_version","alias_value":"2607.28457v1","created_at":"2026-07-31T01:37:40.757977+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.28457","created_at":"2026-07-31T01:37:40.757977+00:00"},{"alias_kind":"pith_short_12","alias_value":"SG6WVWVPS76O","created_at":"2026-07-31T01:37:40.757977+00:00"},{"alias_kind":"pith_short_16","alias_value":"SG6WVWVPS76OFAEK","created_at":"2026-07-31T01:37:40.757977+00:00"},{"alias_kind":"pith_short_8","alias_value":"SG6WVWVP","created_at":"2026-07-31T01:37:40.757977+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/SG6WVWVPS76OFAEKG36XENOW4R","json":"https://pith.science/pith/SG6WVWVPS76OFAEKG36XENOW4R.json","graph_json":"https://pith.science/api/pith-number/SG6WVWVPS76OFAEKG36XENOW4R/graph.json","events_json":"https://pith.science/api/pith-number/SG6WVWVPS76OFAEKG36XENOW4R/events.json","paper":"https://pith.science/paper/SG6WVWVP"},"agent_actions":{"view_html":"https://pith.science/pith/SG6WVWVPS76OFAEKG36XENOW4R","download_json":"https://pith.science/pith/SG6WVWVPS76OFAEKG36XENOW4R.json","view_paper":"https://pith.science/paper/SG6WVWVP","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2607.28457&json=true","fetch_graph":"https://pith.science/api/pith-number/SG6WVWVPS76OFAEKG36XENOW4R/graph.json","fetch_events":"https://pith.science/api/pith-number/SG6WVWVPS76OFAEKG36XENOW4R/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/SG6WVWVPS76OFAEKG36XENOW4R/action/timestamp_anchor","attest_storage":"https://pith.science/pith/SG6WVWVPS76OFAEKG36XENOW4R/action/storage_attestation","attest_author":"https://pith.science/pith/SG6WVWVPS76OFAEKG36XENOW4R/action/author_attestation","sign_citation":"https://pith.science/pith/SG6WVWVPS76OFAEKG36XENOW4R/action/citation_signature","submit_replication":"https://pith.science/pith/SG6WVWVPS76OFAEKG36XENOW4R/action/replication_record"}},"created_at":"2026-07-31T01:37:40.757977+00:00","updated_at":"2026-07-31T01:37:40.757977+00:00"}