{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:O2K7OBCSJIORKYW3HR65WW7NG6","short_pith_number":"pith:O2K7OBCS","schema_version":"1.0","canonical_sha256":"7695f704524a1d1562db3c7ddb5bed378e6e3906404f3940acfb730d9bba7797","source":{"kind":"arxiv","id":"2606.30852","version":1},"attestation_state":"computed","paper":{"title":"When Does Learning to Stop Help? A Cost-Aware Study of Early Exits in Reasoning Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CL","cs.LG"],"primary_cat":"cs.AI","authors_text":"Fang Qin (Stanford University), Manish Shah (Independent Researcher), Zhe Dong (University of Maine at Presque Isle)","submitted_at":"2026-06-29T19:33:42Z","abstract_excerpt":"Reasoning models spend different amounts of useful computation across instances, but it remains unclear when a learned stopping rule improves over simple confidence or convergence thresholds. We study this question with LearnStop, a hidden-state-free checkpoint stopper for reasoning language models. At fixed budget checkpoints, LearnStop probes a short answer from the current reasoning prefix and predicts prefix correctness from online features such as answer confidence, entropy, prefix vote share, answer stability, and backtracking-marker density. Across 18 task-model settings spanning GSM8K,"},"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":"2606.30852","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2026-06-29T19:33:42Z","cross_cats_sorted":["cs.CL","cs.LG"],"title_canon_sha256":"fe16fe075b038791f5b60e516f70068e42381e348cb2d80b65c464982f0f2316","abstract_canon_sha256":"95c03846630cc1ca3c84cb1d9662a8b99c4c7da827d62f8cb9f2080fab58ea02"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-01T00:17:19.500598Z","signature_b64":"qe5CrJ7LaulDjUra+ziVPXqpt0Bbm2DiTuBixoLZSD090WEdfFhkX/v9qk3Qt/MsbuWxxwr/bdOyV3VnZ5/+CA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"7695f704524a1d1562db3c7ddb5bed378e6e3906404f3940acfb730d9bba7797","last_reissued_at":"2026-07-01T00:17:19.500187Z","signature_status":"signed_v1","first_computed_at":"2026-07-01T00:17:19.500187Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"When Does Learning to Stop Help? A Cost-Aware Study of Early Exits in Reasoning Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CL","cs.LG"],"primary_cat":"cs.AI","authors_text":"Fang Qin (Stanford University), Manish Shah (Independent Researcher), Zhe Dong (University of Maine at Presque Isle)","submitted_at":"2026-06-29T19:33:42Z","abstract_excerpt":"Reasoning models spend different amounts of useful computation across instances, but it remains unclear when a learned stopping rule improves over simple confidence or convergence thresholds. We study this question with LearnStop, a hidden-state-free checkpoint stopper for reasoning language models. At fixed budget checkpoints, LearnStop probes a short answer from the current reasoning prefix and predicts prefix correctness from online features such as answer confidence, entropy, prefix vote share, answer stability, and backtracking-marker density. Across 18 task-model settings spanning GSM8K,"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.30852","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/2606.30852/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":"2606.30852","created_at":"2026-07-01T00:17:19.500245+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.30852v1","created_at":"2026-07-01T00:17:19.500245+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.30852","created_at":"2026-07-01T00:17:19.500245+00:00"},{"alias_kind":"pith_short_12","alias_value":"O2K7OBCSJIOR","created_at":"2026-07-01T00:17:19.500245+00:00"},{"alias_kind":"pith_short_16","alias_value":"O2K7OBCSJIORKYW3","created_at":"2026-07-01T00:17:19.500245+00:00"},{"alias_kind":"pith_short_8","alias_value":"O2K7OBCS","created_at":"2026-07-01T00:17:19.500245+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/O2K7OBCSJIORKYW3HR65WW7NG6","json":"https://pith.science/pith/O2K7OBCSJIORKYW3HR65WW7NG6.json","graph_json":"https://pith.science/api/pith-number/O2K7OBCSJIORKYW3HR65WW7NG6/graph.json","events_json":"https://pith.science/api/pith-number/O2K7OBCSJIORKYW3HR65WW7NG6/events.json","paper":"https://pith.science/paper/O2K7OBCS"},"agent_actions":{"view_html":"https://pith.science/pith/O2K7OBCSJIORKYW3HR65WW7NG6","download_json":"https://pith.science/pith/O2K7OBCSJIORKYW3HR65WW7NG6.json","view_paper":"https://pith.science/paper/O2K7OBCS","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.30852&json=true","fetch_graph":"https://pith.science/api/pith-number/O2K7OBCSJIORKYW3HR65WW7NG6/graph.json","fetch_events":"https://pith.science/api/pith-number/O2K7OBCSJIORKYW3HR65WW7NG6/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/O2K7OBCSJIORKYW3HR65WW7NG6/action/timestamp_anchor","attest_storage":"https://pith.science/pith/O2K7OBCSJIORKYW3HR65WW7NG6/action/storage_attestation","attest_author":"https://pith.science/pith/O2K7OBCSJIORKYW3HR65WW7NG6/action/author_attestation","sign_citation":"https://pith.science/pith/O2K7OBCSJIORKYW3HR65WW7NG6/action/citation_signature","submit_replication":"https://pith.science/pith/O2K7OBCSJIORKYW3HR65WW7NG6/action/replication_record"}},"created_at":"2026-07-01T00:17:19.500245+00:00","updated_at":"2026-07-01T00:17:19.500245+00:00"}