{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:5QVHR4QMBWYXQGPQSLPF2M2VIX","short_pith_number":"pith:5QVHR4QM","schema_version":"1.0","canonical_sha256":"ec2a78f20c0db17819f092de5d335545f521dbbd052780c9c1208888f6af99ef","source":{"kind":"arxiv","id":"2310.05910","version":2},"attestation_state":"computed","paper":{"title":"SALMON: Self-Alignment with Instructable Reward Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Chuang Gan, David Cox, Hongxin Zhang, Qinhong Zhou, Yikang Shen, Yiming Yang, Zhenfang Chen, Zhiqing Sun","submitted_at":"2023-10-09T17:56:53Z","abstract_excerpt":"Supervised Fine-Tuning (SFT) on response demonstrations combined with Reinforcement Learning from Human Feedback (RLHF) constitutes a powerful paradigm for aligning LLM-based AI agents. However, a significant limitation of such an approach is its dependency on high-quality human annotations, making its application to intricate tasks challenging due to difficulties in obtaining consistent response demonstrations and in-distribution response preferences. This paper presents a novel approach, namely SALMON, to align base language models with minimal human supervision, using only a small set of hu"},"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":"2310.05910","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2023-10-09T17:56:53Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"a2dd79751ef39d8cdb97201cd18c25348d7fd1c55b6ff7b08cd985135a36d309","abstract_canon_sha256":"86fa098a7239ab6c351d0f94eb07e76ccc15c6c4172bd17362b7a08b2df41b00"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:06:11.662891Z","signature_b64":"7qgk5OGjt8XPUiwN41tc+cf9w4FpNZ6glkE2PlLqn0LLkIRGLwB8G2UQLBZRYaY4mfUksxSFk5CfRcPqLi74Bg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"ec2a78f20c0db17819f092de5d335545f521dbbd052780c9c1208888f6af99ef","last_reissued_at":"2026-07-05T08:06:11.662386Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:06:11.662386Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"SALMON: Self-Alignment with Instructable Reward Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Chuang Gan, David Cox, Hongxin Zhang, Qinhong Zhou, Yikang Shen, Yiming Yang, Zhenfang Chen, Zhiqing Sun","submitted_at":"2023-10-09T17:56:53Z","abstract_excerpt":"Supervised Fine-Tuning (SFT) on response demonstrations combined with Reinforcement Learning from Human Feedback (RLHF) constitutes a powerful paradigm for aligning LLM-based AI agents. However, a significant limitation of such an approach is its dependency on high-quality human annotations, making its application to intricate tasks challenging due to difficulties in obtaining consistent response demonstrations and in-distribution response preferences. This paper presents a novel approach, namely SALMON, to align base language models with minimal human supervision, using only a small set of hu"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2310.05910","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/2310.05910/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":"2310.05910","created_at":"2026-07-05T08:06:11.662454+00:00"},{"alias_kind":"arxiv_version","alias_value":"2310.05910v2","created_at":"2026-07-05T08:06:11.662454+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2310.05910","created_at":"2026-07-05T08:06:11.662454+00:00"},{"alias_kind":"pith_short_12","alias_value":"5QVHR4QMBWYX","created_at":"2026-07-05T08:06:11.662454+00:00"},{"alias_kind":"pith_short_16","alias_value":"5QVHR4QMBWYXQGPQ","created_at":"2026-07-05T08:06:11.662454+00:00"},{"alias_kind":"pith_short_8","alias_value":"5QVHR4QM","created_at":"2026-07-05T08:06:11.662454+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2604.16335","citing_title":"Beyond Verifiable Rewards: Rubric-Based GRM for Reinforced Fine-Tuning SWE Agents","ref_index":24,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/5QVHR4QMBWYXQGPQSLPF2M2VIX","json":"https://pith.science/pith/5QVHR4QMBWYXQGPQSLPF2M2VIX.json","graph_json":"https://pith.science/api/pith-number/5QVHR4QMBWYXQGPQSLPF2M2VIX/graph.json","events_json":"https://pith.science/api/pith-number/5QVHR4QMBWYXQGPQSLPF2M2VIX/events.json","paper":"https://pith.science/paper/5QVHR4QM"},"agent_actions":{"view_html":"https://pith.science/pith/5QVHR4QMBWYXQGPQSLPF2M2VIX","download_json":"https://pith.science/pith/5QVHR4QMBWYXQGPQSLPF2M2VIX.json","view_paper":"https://pith.science/paper/5QVHR4QM","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2310.05910&json=true","fetch_graph":"https://pith.science/api/pith-number/5QVHR4QMBWYXQGPQSLPF2M2VIX/graph.json","fetch_events":"https://pith.science/api/pith-number/5QVHR4QMBWYXQGPQSLPF2M2VIX/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/5QVHR4QMBWYXQGPQSLPF2M2VIX/action/timestamp_anchor","attest_storage":"https://pith.science/pith/5QVHR4QMBWYXQGPQSLPF2M2VIX/action/storage_attestation","attest_author":"https://pith.science/pith/5QVHR4QMBWYXQGPQSLPF2M2VIX/action/author_attestation","sign_citation":"https://pith.science/pith/5QVHR4QMBWYXQGPQSLPF2M2VIX/action/citation_signature","submit_replication":"https://pith.science/pith/5QVHR4QMBWYXQGPQSLPF2M2VIX/action/replication_record"}},"created_at":"2026-07-05T08:06:11.662454+00:00","updated_at":"2026-07-05T08:06:11.662454+00:00"}