{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:KYNOXZXFLYU2A4ZDCWSKCKOGXJ","short_pith_number":"pith:KYNOXZXF","schema_version":"1.0","canonical_sha256":"561aebe6e55e29a0732315a4a129c6ba6df715b31f458539883954bd5f98b799","source":{"kind":"arxiv","id":"2406.16897","version":1},"attestation_state":"computed","paper":{"title":"InstructPatentGPT: Training patent language models to follow instructions with human feedback","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Jieh-Sheng Lee","submitted_at":"2024-05-25T11:48:50Z","abstract_excerpt":"In this research, patent prosecution is conceptualized as a system of reinforcement learning from human feedback. The objective of the system is to increase the likelihood for a language model to generate patent claims that have a higher chance of being granted. To showcase the controllability of the language model, the system learns from granted patents and pre-grant applications with different rewards. The status of \"granted\" and \"pre-grant\" are perceived as labeled human feedback implicitly. In addition, specific to patent drafting, the experiments in this research demonstrate the model's c"},"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":"2406.16897","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-05-25T11:48:50Z","cross_cats_sorted":[],"title_canon_sha256":"92a9c15b7119b8d9b54da57a14a04383382e5c41fe5493df7f10ba30f200dee4","abstract_canon_sha256":"906b26b4d702b82eea6f72fab68786711c825d9b771ee102d897961b69402c2e"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:36:06.623753Z","signature_b64":"+kB0fPG55nZTyahgvrwfH0pwc7YZK/1V0WZy9BGvie+7rHYG8SOw6sm8d3o/BxsNsxuBWLh1qeMpi6rqfWBhAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"561aebe6e55e29a0732315a4a129c6ba6df715b31f458539883954bd5f98b799","last_reissued_at":"2026-07-05T08:36:06.623354Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:36:06.623354Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"InstructPatentGPT: Training patent language models to follow instructions with human feedback","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Jieh-Sheng Lee","submitted_at":"2024-05-25T11:48:50Z","abstract_excerpt":"In this research, patent prosecution is conceptualized as a system of reinforcement learning from human feedback. The objective of the system is to increase the likelihood for a language model to generate patent claims that have a higher chance of being granted. To showcase the controllability of the language model, the system learns from granted patents and pre-grant applications with different rewards. The status of \"granted\" and \"pre-grant\" are perceived as labeled human feedback implicitly. In addition, specific to patent drafting, the experiments in this research demonstrate the model's c"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2406.16897","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/2406.16897/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":"2406.16897","created_at":"2026-07-05T08:36:06.623409+00:00"},{"alias_kind":"arxiv_version","alias_value":"2406.16897v1","created_at":"2026-07-05T08:36:06.623409+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2406.16897","created_at":"2026-07-05T08:36:06.623409+00:00"},{"alias_kind":"pith_short_12","alias_value":"KYNOXZXFLYU2","created_at":"2026-07-05T08:36:06.623409+00:00"},{"alias_kind":"pith_short_16","alias_value":"KYNOXZXFLYU2A4ZD","created_at":"2026-07-05T08:36:06.623409+00:00"},{"alias_kind":"pith_short_8","alias_value":"KYNOXZXF","created_at":"2026-07-05T08:36:06.623409+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2507.16951","citing_title":"Harnessing RLHF for Robust Unanswerability Recognition and Trustworthy Response Generation in LLMs","ref_index":7,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/KYNOXZXFLYU2A4ZDCWSKCKOGXJ","json":"https://pith.science/pith/KYNOXZXFLYU2A4ZDCWSKCKOGXJ.json","graph_json":"https://pith.science/api/pith-number/KYNOXZXFLYU2A4ZDCWSKCKOGXJ/graph.json","events_json":"https://pith.science/api/pith-number/KYNOXZXFLYU2A4ZDCWSKCKOGXJ/events.json","paper":"https://pith.science/paper/KYNOXZXF"},"agent_actions":{"view_html":"https://pith.science/pith/KYNOXZXFLYU2A4ZDCWSKCKOGXJ","download_json":"https://pith.science/pith/KYNOXZXFLYU2A4ZDCWSKCKOGXJ.json","view_paper":"https://pith.science/paper/KYNOXZXF","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2406.16897&json=true","fetch_graph":"https://pith.science/api/pith-number/KYNOXZXFLYU2A4ZDCWSKCKOGXJ/graph.json","fetch_events":"https://pith.science/api/pith-number/KYNOXZXFLYU2A4ZDCWSKCKOGXJ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/KYNOXZXFLYU2A4ZDCWSKCKOGXJ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/KYNOXZXFLYU2A4ZDCWSKCKOGXJ/action/storage_attestation","attest_author":"https://pith.science/pith/KYNOXZXFLYU2A4ZDCWSKCKOGXJ/action/author_attestation","sign_citation":"https://pith.science/pith/KYNOXZXFLYU2A4ZDCWSKCKOGXJ/action/citation_signature","submit_replication":"https://pith.science/pith/KYNOXZXFLYU2A4ZDCWSKCKOGXJ/action/replication_record"}},"created_at":"2026-07-05T08:36:06.623409+00:00","updated_at":"2026-07-05T08:36:06.623409+00:00"}