{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:JXVYNV2NNPXWMHPLKZ2YQFWAUX","short_pith_number":"pith:JXVYNV2N","schema_version":"1.0","canonical_sha256":"4deb86d74d6bef661deb56758816c0a5d2cbe0e24e29ca18f84a5c035f26388a","source":{"kind":"arxiv","id":"2405.19332","version":3},"attestation_state":"computed","paper":{"title":"Self-Exploring Language Models: Active Preference Elicitation for Online Alignment","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Donghan Yu, Hany Hassan, Han Zhong, Hiteshi Sharma, Shenao Zhang, Shuohang Wang, Zhaoran Wang, Zhihan Liu, Ziyi Yang","submitted_at":"2024-05-29T17:59:07Z","abstract_excerpt":"Preference optimization, particularly through Reinforcement Learning from Human Feedback (RLHF), has achieved significant success in aligning Large Language Models (LLMs) to adhere to human intentions. Unlike offline alignment with a fixed dataset, online feedback collection from humans or AI on model generations typically leads to more capable reward models and better-aligned LLMs through an iterative process. However, achieving a globally accurate reward model requires systematic exploration to generate diverse responses that span the vast space of natural language. Random sampling from stan"},"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":"2405.19332","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-05-29T17:59:07Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"8f723b524f9cb686809dcdecfce5c6b2634b217bbcbeae901f6810a8c6345fce","abstract_canon_sha256":"b39e3fb4575922279bd512f437e97f30741576820e746abcbdf596a424e5584f"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:31:03.277699Z","signature_b64":"aEz8+YlxqazQE/hke0u7DZp7bvQCRHqgBYX1nFCFw4npZ5gYRqK2qJbeK8Hyy9lJw6L94oAojQlvxaEsKcnfAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"4deb86d74d6bef661deb56758816c0a5d2cbe0e24e29ca18f84a5c035f26388a","last_reissued_at":"2026-07-05T09:31:03.277151Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:31:03.277151Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Self-Exploring Language Models: Active Preference Elicitation for Online Alignment","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Donghan Yu, Hany Hassan, Han Zhong, Hiteshi Sharma, Shenao Zhang, Shuohang Wang, Zhaoran Wang, Zhihan Liu, Ziyi Yang","submitted_at":"2024-05-29T17:59:07Z","abstract_excerpt":"Preference optimization, particularly through Reinforcement Learning from Human Feedback (RLHF), has achieved significant success in aligning Large Language Models (LLMs) to adhere to human intentions. Unlike offline alignment with a fixed dataset, online feedback collection from humans or AI on model generations typically leads to more capable reward models and better-aligned LLMs through an iterative process. However, achieving a globally accurate reward model requires systematic exploration to generate diverse responses that span the vast space of natural language. Random sampling from stan"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2405.19332","kind":"arxiv","version":3},"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/2405.19332/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":"2405.19332","created_at":"2026-07-05T09:31:03.277223+00:00"},{"alias_kind":"arxiv_version","alias_value":"2405.19332v3","created_at":"2026-07-05T09:31:03.277223+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2405.19332","created_at":"2026-07-05T09:31:03.277223+00:00"},{"alias_kind":"pith_short_12","alias_value":"JXVYNV2NNPXW","created_at":"2026-07-05T09:31:03.277223+00:00"},{"alias_kind":"pith_short_16","alias_value":"JXVYNV2NNPXWMHPL","created_at":"2026-07-05T09:31:03.277223+00:00"},{"alias_kind":"pith_short_8","alias_value":"JXVYNV2N","created_at":"2026-07-05T09:31:03.277223+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":4,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.04477","citing_title":"Data-dependent Exploration for Online Reinforcement Learning from Human Feedback","ref_index":36,"is_internal_anchor":false},{"citing_arxiv_id":"2605.20408","citing_title":"Spectral Souping: A Unified Framework for Online Preference Alignment","ref_index":36,"is_internal_anchor":false},{"citing_arxiv_id":"2605.16712","citing_title":"Recall Isn't Enough: Bounding Commitments in Personalized Language Systems","ref_index":12,"is_internal_anchor":false},{"citing_arxiv_id":"2605.04477","citing_title":"Data-dependent Exploration for Online Reinforcement Learning from Human Feedback","ref_index":90,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/JXVYNV2NNPXWMHPLKZ2YQFWAUX","json":"https://pith.science/pith/JXVYNV2NNPXWMHPLKZ2YQFWAUX.json","graph_json":"https://pith.science/api/pith-number/JXVYNV2NNPXWMHPLKZ2YQFWAUX/graph.json","events_json":"https://pith.science/api/pith-number/JXVYNV2NNPXWMHPLKZ2YQFWAUX/events.json","paper":"https://pith.science/paper/JXVYNV2N"},"agent_actions":{"view_html":"https://pith.science/pith/JXVYNV2NNPXWMHPLKZ2YQFWAUX","download_json":"https://pith.science/pith/JXVYNV2NNPXWMHPLKZ2YQFWAUX.json","view_paper":"https://pith.science/paper/JXVYNV2N","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2405.19332&json=true","fetch_graph":"https://pith.science/api/pith-number/JXVYNV2NNPXWMHPLKZ2YQFWAUX/graph.json","fetch_events":"https://pith.science/api/pith-number/JXVYNV2NNPXWMHPLKZ2YQFWAUX/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/JXVYNV2NNPXWMHPLKZ2YQFWAUX/action/timestamp_anchor","attest_storage":"https://pith.science/pith/JXVYNV2NNPXWMHPLKZ2YQFWAUX/action/storage_attestation","attest_author":"https://pith.science/pith/JXVYNV2NNPXWMHPLKZ2YQFWAUX/action/author_attestation","sign_citation":"https://pith.science/pith/JXVYNV2NNPXWMHPLKZ2YQFWAUX/action/citation_signature","submit_replication":"https://pith.science/pith/JXVYNV2NNPXWMHPLKZ2YQFWAUX/action/replication_record"}},"created_at":"2026-07-05T09:31:03.277223+00:00","updated_at":"2026-07-05T09:31:03.277223+00:00"}