{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:KAGSCWG7LBD54WVIEUYFAUUN5E","short_pith_number":"pith:KAGSCWG7","schema_version":"1.0","canonical_sha256":"500d2158df5847de5aa8253050528de921ee3aa5003375f9aaa84b826e55cff8","source":{"kind":"arxiv","id":"2608.08020","version":1},"attestation_state":"computed","paper":{"title":"Thought-Level Beam Search for Reasoning","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Hongyin Luo, Lijie Yang, Ravi Netravali, Tri Dao","submitted_at":"2026-08-08T09:00:01Z","abstract_excerpt":"Test-time compute scaling is a primary driver of performance in large reasoning models (LRMs), but extreme inefficiency bounds current approaches, shifting the critical question from \\emph{how much} compute to spend, to \\emph{where} to allocate it. We formalize test-time reasoning as a constrained compute allocation problem over partial trajectories. Under a fixed hardware budget, existing paradigms fail to actively allocate the compute to the most promising partial progress: traditional parallel sampling treats traces independently and induces severe memory bottlenecks, while subtractive prun"},"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":"2608.08020","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.AI","submitted_at":"2026-08-08T09:00:01Z","cross_cats_sorted":[],"title_canon_sha256":"85c64d2f9ba77d23dbba6f2a1650a3844b2c363c8d38c1836f5458ba31565b68","abstract_canon_sha256":"b1fa18f93275e32ad97aa8b3034f5e95fb2b8a5196abcaeea12757e35381e9ee"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-08-11T01:21:25.166083Z","signature_b64":"9MfDFfpZYqvz0cRqyJ1JTSQc+P4DeU/5yVzIC+rPJW6SPORpALz1rXJb1geWcIHjmsApjHxRDn8eccB5+tlECw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"500d2158df5847de5aa8253050528de921ee3aa5003375f9aaa84b826e55cff8","last_reissued_at":"2026-08-11T01:21:25.163493Z","signature_status":"signed_v1","first_computed_at":"2026-08-11T01:21:25.163493Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Thought-Level Beam Search for Reasoning","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Hongyin Luo, Lijie Yang, Ravi Netravali, Tri Dao","submitted_at":"2026-08-08T09:00:01Z","abstract_excerpt":"Test-time compute scaling is a primary driver of performance in large reasoning models (LRMs), but extreme inefficiency bounds current approaches, shifting the critical question from \\emph{how much} compute to spend, to \\emph{where} to allocate it. We formalize test-time reasoning as a constrained compute allocation problem over partial trajectories. Under a fixed hardware budget, existing paradigms fail to actively allocate the compute to the most promising partial progress: traditional parallel sampling treats traces independently and induces severe memory bottlenecks, while subtractive prun"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2608.08020","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/2608.08020/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":"2608.08020","created_at":"2026-08-11T01:21:25.164278+00:00"},{"alias_kind":"arxiv_version","alias_value":"2608.08020v1","created_at":"2026-08-11T01:21:25.164278+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2608.08020","created_at":"2026-08-11T01:21:25.164278+00:00"},{"alias_kind":"pith_short_12","alias_value":"KAGSCWG7LBD5","created_at":"2026-08-11T01:21:25.164278+00:00"},{"alias_kind":"pith_short_16","alias_value":"KAGSCWG7LBD54WVI","created_at":"2026-08-11T01:21:25.164278+00:00"},{"alias_kind":"pith_short_8","alias_value":"KAGSCWG7","created_at":"2026-08-11T01:21:25.164278+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/KAGSCWG7LBD54WVIEUYFAUUN5E","json":"https://pith.science/pith/KAGSCWG7LBD54WVIEUYFAUUN5E.json","graph_json":"https://pith.science/api/pith-number/KAGSCWG7LBD54WVIEUYFAUUN5E/graph.json","events_json":"https://pith.science/api/pith-number/KAGSCWG7LBD54WVIEUYFAUUN5E/events.json","paper":"https://pith.science/paper/KAGSCWG7"},"agent_actions":{"view_html":"https://pith.science/pith/KAGSCWG7LBD54WVIEUYFAUUN5E","download_json":"https://pith.science/pith/KAGSCWG7LBD54WVIEUYFAUUN5E.json","view_paper":"https://pith.science/paper/KAGSCWG7","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2608.08020&json=true","fetch_graph":"https://pith.science/api/pith-number/KAGSCWG7LBD54WVIEUYFAUUN5E/graph.json","fetch_events":"https://pith.science/api/pith-number/KAGSCWG7LBD54WVIEUYFAUUN5E/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/KAGSCWG7LBD54WVIEUYFAUUN5E/action/timestamp_anchor","attest_storage":"https://pith.science/pith/KAGSCWG7LBD54WVIEUYFAUUN5E/action/storage_attestation","attest_author":"https://pith.science/pith/KAGSCWG7LBD54WVIEUYFAUUN5E/action/author_attestation","sign_citation":"https://pith.science/pith/KAGSCWG7LBD54WVIEUYFAUUN5E/action/citation_signature","submit_replication":"https://pith.science/pith/KAGSCWG7LBD54WVIEUYFAUUN5E/action/replication_record"}},"created_at":"2026-08-11T01:21:25.164278+00:00","updated_at":"2026-08-11T01:21:25.164278+00:00"}