{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:GGN5JAISBHGMFBIOIFK3R2WFEL","short_pith_number":"pith:GGN5JAIS","schema_version":"1.0","canonical_sha256":"319bd4811209ccc2850e4155b8eac522d8094d1cba9928f0bddba95f54543b04","source":{"kind":"arxiv","id":"2507.09016","version":2},"attestation_state":"computed","paper":{"title":"Enhancing RLHF with Human Gaze Modeling","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Ilya Pershin, Ivan Titov, Karim Galliamov","submitted_at":"2025-07-11T20:49:04Z","abstract_excerpt":"Reinforcement Learning from Human Feedback (RLHF) aligns language models with human preferences but is computationally expensive. We explore two approaches that leverage human gaze modeling to enhance RLHF: (1) gaze-aware reward models and (2) gaze-based distribution of sparse rewards at token level. Our experiments demonstate that gaze-informed RLHF achieves faster convergence while maintaining or slightly improving performance, thus, reducing computational costs during policy optimization. These results show that human gaze provides a valuable and underused signal for policy optimization, po"},"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":"2507.09016","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.LG","submitted_at":"2025-07-11T20:49:04Z","cross_cats_sorted":[],"title_canon_sha256":"a866a9373f9417976fc3c292a913b8721c9f0fb05fcfa219bf7fa9d3052d6c58","abstract_canon_sha256":"78373312a58171816553fb87f0d5f81fc94d0b2aa4043cca210b9b1718319900"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:38:09.173408Z","signature_b64":"AO0RF5iP3TuZSOpnoyw4CXwA6yhNb2gMgKR3VPhFYRfwMGCX2G2xu65sIOOi83U4YyJVxyDBCA7F3WacRP8RBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"319bd4811209ccc2850e4155b8eac522d8094d1cba9928f0bddba95f54543b04","last_reissued_at":"2026-07-05T11:38:09.172888Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:38:09.172888Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Enhancing RLHF with Human Gaze Modeling","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Ilya Pershin, Ivan Titov, Karim Galliamov","submitted_at":"2025-07-11T20:49:04Z","abstract_excerpt":"Reinforcement Learning from Human Feedback (RLHF) aligns language models with human preferences but is computationally expensive. We explore two approaches that leverage human gaze modeling to enhance RLHF: (1) gaze-aware reward models and (2) gaze-based distribution of sparse rewards at token level. Our experiments demonstate that gaze-informed RLHF achieves faster convergence while maintaining or slightly improving performance, thus, reducing computational costs during policy optimization. These results show that human gaze provides a valuable and underused signal for policy optimization, po"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2507.09016","kind":"arxiv","version":2},"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/2507.09016/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":"2507.09016","created_at":"2026-07-05T11:38:09.172950+00:00"},{"alias_kind":"arxiv_version","alias_value":"2507.09016v2","created_at":"2026-07-05T11:38:09.172950+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2507.09016","created_at":"2026-07-05T11:38:09.172950+00:00"},{"alias_kind":"pith_short_12","alias_value":"GGN5JAISBHGM","created_at":"2026-07-05T11:38:09.172950+00:00"},{"alias_kind":"pith_short_16","alias_value":"GGN5JAISBHGMFBIO","created_at":"2026-07-05T11:38:09.172950+00:00"},{"alias_kind":"pith_short_8","alias_value":"GGN5JAIS","created_at":"2026-07-05T11:38:09.172950+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/GGN5JAISBHGMFBIOIFK3R2WFEL","json":"https://pith.science/pith/GGN5JAISBHGMFBIOIFK3R2WFEL.json","graph_json":"https://pith.science/api/pith-number/GGN5JAISBHGMFBIOIFK3R2WFEL/graph.json","events_json":"https://pith.science/api/pith-number/GGN5JAISBHGMFBIOIFK3R2WFEL/events.json","paper":"https://pith.science/paper/GGN5JAIS"},"agent_actions":{"view_html":"https://pith.science/pith/GGN5JAISBHGMFBIOIFK3R2WFEL","download_json":"https://pith.science/pith/GGN5JAISBHGMFBIOIFK3R2WFEL.json","view_paper":"https://pith.science/paper/GGN5JAIS","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2507.09016&json=true","fetch_graph":"https://pith.science/api/pith-number/GGN5JAISBHGMFBIOIFK3R2WFEL/graph.json","fetch_events":"https://pith.science/api/pith-number/GGN5JAISBHGMFBIOIFK3R2WFEL/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/GGN5JAISBHGMFBIOIFK3R2WFEL/action/timestamp_anchor","attest_storage":"https://pith.science/pith/GGN5JAISBHGMFBIOIFK3R2WFEL/action/storage_attestation","attest_author":"https://pith.science/pith/GGN5JAISBHGMFBIOIFK3R2WFEL/action/author_attestation","sign_citation":"https://pith.science/pith/GGN5JAISBHGMFBIOIFK3R2WFEL/action/citation_signature","submit_replication":"https://pith.science/pith/GGN5JAISBHGMFBIOIFK3R2WFEL/action/replication_record"}},"created_at":"2026-07-05T11:38:09.172950+00:00","updated_at":"2026-07-05T11:38:09.172950+00:00"}