{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:SV2CP7DQZPIAV4WZPFHBFEC5KZ","short_pith_number":"pith:SV2CP7DQ","schema_version":"1.0","canonical_sha256":"957427fc70cbd00af2d9794e12905d564d8e615f65ae50be23fc857c0348523d","source":{"kind":"arxiv","id":"2410.08048","version":1},"attestation_state":"computed","paper":{"title":"VerifierQ: Enhancing LLM Test Time Compute with Q-Learning-based Verifiers","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.LG","authors_text":"Hao Tang, Jianing Qi, Zhigang Zhu","submitted_at":"2024-10-10T15:43:55Z","abstract_excerpt":"Recent advancements in test time compute, particularly through the use of verifier models, have significantly enhanced the reasoning capabilities of Large Language Models (LLMs). This generator-verifier approach closely resembles the actor-critic framework in reinforcement learning (RL). However, current verifier models in LLMs often rely on supervised fine-tuning without temporal difference learning such as Q-learning. This paper introduces VerifierQ, a novel approach that integrates Offline Q-learning into LLM verifier models. We address three key challenges in applying Q-learning to LLMs: ("},"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":"2410.08048","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-10-10T15:43:55Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"66a6d37c918b2b05c534fd2c38c63bdc1b449f3d4169abde9783f32e14464c3e","abstract_canon_sha256":"a6736a93b8acf07c58ec1fdf93f414deac667fc9390bed72286533d1d8489b7e"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:18:47.873051Z","signature_b64":"0RYnLkHBmp6Kbk8EAAaWTk64eKxCIjBj0RlQovBJ01VVb+ugtYNfqNCXFTPVOe4gAH4I7HHm5QyIbnqJp0IKBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"957427fc70cbd00af2d9794e12905d564d8e615f65ae50be23fc857c0348523d","last_reissued_at":"2026-07-05T09:18:47.872568Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:18:47.872568Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"VerifierQ: Enhancing LLM Test Time Compute with Q-Learning-based Verifiers","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.LG","authors_text":"Hao Tang, Jianing Qi, Zhigang Zhu","submitted_at":"2024-10-10T15:43:55Z","abstract_excerpt":"Recent advancements in test time compute, particularly through the use of verifier models, have significantly enhanced the reasoning capabilities of Large Language Models (LLMs). This generator-verifier approach closely resembles the actor-critic framework in reinforcement learning (RL). However, current verifier models in LLMs often rely on supervised fine-tuning without temporal difference learning such as Q-learning. This paper introduces VerifierQ, a novel approach that integrates Offline Q-learning into LLM verifier models. We address three key challenges in applying Q-learning to LLMs: ("},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.08048","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/2410.08048/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":"2410.08048","created_at":"2026-07-05T09:18:47.872624+00:00"},{"alias_kind":"arxiv_version","alias_value":"2410.08048v1","created_at":"2026-07-05T09:18:47.872624+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.08048","created_at":"2026-07-05T09:18:47.872624+00:00"},{"alias_kind":"pith_short_12","alias_value":"SV2CP7DQZPIA","created_at":"2026-07-05T09:18:47.872624+00:00"},{"alias_kind":"pith_short_16","alias_value":"SV2CP7DQZPIAV4WZ","created_at":"2026-07-05T09:18:47.872624+00:00"},{"alias_kind":"pith_short_8","alias_value":"SV2CP7DQ","created_at":"2026-07-05T09:18:47.872624+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2512.02080","citing_title":"The 4/$\\delta$ Bound: Designing Predictable LLM-Verifier Systems for Formal Method Guarantee","ref_index":18,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/SV2CP7DQZPIAV4WZPFHBFEC5KZ","json":"https://pith.science/pith/SV2CP7DQZPIAV4WZPFHBFEC5KZ.json","graph_json":"https://pith.science/api/pith-number/SV2CP7DQZPIAV4WZPFHBFEC5KZ/graph.json","events_json":"https://pith.science/api/pith-number/SV2CP7DQZPIAV4WZPFHBFEC5KZ/events.json","paper":"https://pith.science/paper/SV2CP7DQ"},"agent_actions":{"view_html":"https://pith.science/pith/SV2CP7DQZPIAV4WZPFHBFEC5KZ","download_json":"https://pith.science/pith/SV2CP7DQZPIAV4WZPFHBFEC5KZ.json","view_paper":"https://pith.science/paper/SV2CP7DQ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2410.08048&json=true","fetch_graph":"https://pith.science/api/pith-number/SV2CP7DQZPIAV4WZPFHBFEC5KZ/graph.json","fetch_events":"https://pith.science/api/pith-number/SV2CP7DQZPIAV4WZPFHBFEC5KZ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/SV2CP7DQZPIAV4WZPFHBFEC5KZ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/SV2CP7DQZPIAV4WZPFHBFEC5KZ/action/storage_attestation","attest_author":"https://pith.science/pith/SV2CP7DQZPIAV4WZPFHBFEC5KZ/action/author_attestation","sign_citation":"https://pith.science/pith/SV2CP7DQZPIAV4WZPFHBFEC5KZ/action/citation_signature","submit_replication":"https://pith.science/pith/SV2CP7DQZPIAV4WZPFHBFEC5KZ/action/replication_record"}},"created_at":"2026-07-05T09:18:47.872624+00:00","updated_at":"2026-07-05T09:18:47.872624+00:00"}