{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:G3PK6RUWIG7T5DRPQOIV2AHU2G","short_pith_number":"pith:G3PK6RUW","schema_version":"1.0","canonical_sha256":"36deaf469641bf3e8e2f83915d00f4d1a94669335d8b837857725012fe660041","source":{"kind":"arxiv","id":"2506.15617","version":1},"attestation_state":"computed","paper":{"title":"The Compositional Architecture of Regret in Large Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Di Wang, Lijie Hu, Shu Yang, Tianjin Huang, Wanyu Lin, Xiangxiang Cui","submitted_at":"2025-06-18T16:50:34Z","abstract_excerpt":"Regret in Large Language Models refers to their explicit regret expression when presented with evidence contradicting their previously generated misinformation. Studying the regret mechanism is crucial for enhancing model reliability and helps in revealing how cognition is coded in neural networks. To understand this mechanism, we need to first identify regret expressions in model outputs, then analyze their internal representation. This analysis requires examining the model's hidden states, where information processing occurs at the neuron level. However, this faces three key challenges: (1) "},"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":"2506.15617","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-06-18T16:50:34Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"a2deae2fc74c6df032717e72af5cfb37ca4e3e034c6345c26a4128f6cf28ec24","abstract_canon_sha256":"d2ff779286bc1fafbc6b9f7884202056d94df9f696ab03dd00d98aadb6a7de21"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:23:47.115750Z","signature_b64":"zlbqrHDtTAUEPwZyQ08ywc6xEILR8K2zW/meBWWZRkw6C4PPjIaH0smBwj0r69rv9ND4ZHt8uyY0xQE5qqGxDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"36deaf469641bf3e8e2f83915d00f4d1a94669335d8b837857725012fe660041","last_reissued_at":"2026-07-05T11:23:47.115123Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:23:47.115123Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"The Compositional Architecture of Regret in Large Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Di Wang, Lijie Hu, Shu Yang, Tianjin Huang, Wanyu Lin, Xiangxiang Cui","submitted_at":"2025-06-18T16:50:34Z","abstract_excerpt":"Regret in Large Language Models refers to their explicit regret expression when presented with evidence contradicting their previously generated misinformation. Studying the regret mechanism is crucial for enhancing model reliability and helps in revealing how cognition is coded in neural networks. To understand this mechanism, we need to first identify regret expressions in model outputs, then analyze their internal representation. This analysis requires examining the model's hidden states, where information processing occurs at the neuron level. However, this faces three key challenges: (1) "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.15617","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/2506.15617/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":"2506.15617","created_at":"2026-07-05T11:23:47.115193+00:00"},{"alias_kind":"arxiv_version","alias_value":"2506.15617v1","created_at":"2026-07-05T11:23:47.115193+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.15617","created_at":"2026-07-05T11:23:47.115193+00:00"},{"alias_kind":"pith_short_12","alias_value":"G3PK6RUWIG7T","created_at":"2026-07-05T11:23:47.115193+00:00"},{"alias_kind":"pith_short_16","alias_value":"G3PK6RUWIG7T5DRP","created_at":"2026-07-05T11:23:47.115193+00:00"},{"alias_kind":"pith_short_8","alias_value":"G3PK6RUW","created_at":"2026-07-05T11:23:47.115193+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/G3PK6RUWIG7T5DRPQOIV2AHU2G","json":"https://pith.science/pith/G3PK6RUWIG7T5DRPQOIV2AHU2G.json","graph_json":"https://pith.science/api/pith-number/G3PK6RUWIG7T5DRPQOIV2AHU2G/graph.json","events_json":"https://pith.science/api/pith-number/G3PK6RUWIG7T5DRPQOIV2AHU2G/events.json","paper":"https://pith.science/paper/G3PK6RUW"},"agent_actions":{"view_html":"https://pith.science/pith/G3PK6RUWIG7T5DRPQOIV2AHU2G","download_json":"https://pith.science/pith/G3PK6RUWIG7T5DRPQOIV2AHU2G.json","view_paper":"https://pith.science/paper/G3PK6RUW","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2506.15617&json=true","fetch_graph":"https://pith.science/api/pith-number/G3PK6RUWIG7T5DRPQOIV2AHU2G/graph.json","fetch_events":"https://pith.science/api/pith-number/G3PK6RUWIG7T5DRPQOIV2AHU2G/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/G3PK6RUWIG7T5DRPQOIV2AHU2G/action/timestamp_anchor","attest_storage":"https://pith.science/pith/G3PK6RUWIG7T5DRPQOIV2AHU2G/action/storage_attestation","attest_author":"https://pith.science/pith/G3PK6RUWIG7T5DRPQOIV2AHU2G/action/author_attestation","sign_citation":"https://pith.science/pith/G3PK6RUWIG7T5DRPQOIV2AHU2G/action/citation_signature","submit_replication":"https://pith.science/pith/G3PK6RUWIG7T5DRPQOIV2AHU2G/action/replication_record"}},"created_at":"2026-07-05T11:23:47.115193+00:00","updated_at":"2026-07-05T11:23:47.115193+00:00"}