{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:RCEF5TZKAH7NX5MT7RBHQGKLEC","short_pith_number":"pith:RCEF5TZK","schema_version":"1.0","canonical_sha256":"88885ecf2a01fedbf593fc4278194b209385827884596f150da98827b1b6ddbb","source":{"kind":"arxiv","id":"2403.09972","version":3},"attestation_state":"computed","paper":{"title":"Think Twice Before Trusting: Self-Detection for Large Language Models through Comprehensive Answer Reflection","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Fengbin Zhu, Fuli Feng, Moxin Li, Qifan Wang, Tat-Seng Chua, Wenjie Wang","submitted_at":"2024-03-15T02:38:26Z","abstract_excerpt":"Self-detection for Large Language Models (LLMs) seeks to evaluate the trustworthiness of the LLM's output by leveraging its own capabilities, thereby alleviating the issue of output hallucination. However, existing self-detection approaches only retrospectively evaluate answers generated by LLM, typically leading to the over-trust in incorrectly generated answers. To tackle this limitation, we propose a novel self-detection paradigm that considers the comprehensive answer space beyond LLM-generated answers. It thoroughly compares the trustworthiness of multiple candidate answers to mitigate th"},"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":"2403.09972","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-03-15T02:38:26Z","cross_cats_sorted":[],"title_canon_sha256":"d7c7edbf22d935db8a3eb0587a7bcc013207ecdab18ed92e92e766fa1ea0a14e","abstract_canon_sha256":"5b799f2b6b7d7282eacf2325d86ecd72aae231865dcce57777f2866bc7e501a0"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:12:26.562384Z","signature_b64":"c/Q1ZhDKFI3LfWxuYHnJtGqs3a9qdsV6D/ns3kHiqsheJOsRULBmRlTIJeeto0cdMN8dUSqbr7jnrgLlO9vnAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"88885ecf2a01fedbf593fc4278194b209385827884596f150da98827b1b6ddbb","last_reissued_at":"2026-07-05T09:12:26.561822Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:12:26.561822Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Think Twice Before Trusting: Self-Detection for Large Language Models through Comprehensive Answer Reflection","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Fengbin Zhu, Fuli Feng, Moxin Li, Qifan Wang, Tat-Seng Chua, Wenjie Wang","submitted_at":"2024-03-15T02:38:26Z","abstract_excerpt":"Self-detection for Large Language Models (LLMs) seeks to evaluate the trustworthiness of the LLM's output by leveraging its own capabilities, thereby alleviating the issue of output hallucination. However, existing self-detection approaches only retrospectively evaluate answers generated by LLM, typically leading to the over-trust in incorrectly generated answers. To tackle this limitation, we propose a novel self-detection paradigm that considers the comprehensive answer space beyond LLM-generated answers. It thoroughly compares the trustworthiness of multiple candidate answers to mitigate th"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2403.09972","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/2403.09972/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":"2403.09972","created_at":"2026-07-05T09:12:26.561881+00:00"},{"alias_kind":"arxiv_version","alias_value":"2403.09972v3","created_at":"2026-07-05T09:12:26.561881+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2403.09972","created_at":"2026-07-05T09:12:26.561881+00:00"},{"alias_kind":"pith_short_12","alias_value":"RCEF5TZKAH7N","created_at":"2026-07-05T09:12:26.561881+00:00"},{"alias_kind":"pith_short_16","alias_value":"RCEF5TZKAH7NX5MT","created_at":"2026-07-05T09:12:26.561881+00:00"},{"alias_kind":"pith_short_8","alias_value":"RCEF5TZK","created_at":"2026-07-05T09:12:26.561881+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2506.16172","citing_title":"SGIC: A Self-Guided Iterative Calibration Framework for RAG","ref_index":22,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/RCEF5TZKAH7NX5MT7RBHQGKLEC","json":"https://pith.science/pith/RCEF5TZKAH7NX5MT7RBHQGKLEC.json","graph_json":"https://pith.science/api/pith-number/RCEF5TZKAH7NX5MT7RBHQGKLEC/graph.json","events_json":"https://pith.science/api/pith-number/RCEF5TZKAH7NX5MT7RBHQGKLEC/events.json","paper":"https://pith.science/paper/RCEF5TZK"},"agent_actions":{"view_html":"https://pith.science/pith/RCEF5TZKAH7NX5MT7RBHQGKLEC","download_json":"https://pith.science/pith/RCEF5TZKAH7NX5MT7RBHQGKLEC.json","view_paper":"https://pith.science/paper/RCEF5TZK","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2403.09972&json=true","fetch_graph":"https://pith.science/api/pith-number/RCEF5TZKAH7NX5MT7RBHQGKLEC/graph.json","fetch_events":"https://pith.science/api/pith-number/RCEF5TZKAH7NX5MT7RBHQGKLEC/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/RCEF5TZKAH7NX5MT7RBHQGKLEC/action/timestamp_anchor","attest_storage":"https://pith.science/pith/RCEF5TZKAH7NX5MT7RBHQGKLEC/action/storage_attestation","attest_author":"https://pith.science/pith/RCEF5TZKAH7NX5MT7RBHQGKLEC/action/author_attestation","sign_citation":"https://pith.science/pith/RCEF5TZKAH7NX5MT7RBHQGKLEC/action/citation_signature","submit_replication":"https://pith.science/pith/RCEF5TZKAH7NX5MT7RBHQGKLEC/action/replication_record"}},"created_at":"2026-07-05T09:12:26.561881+00:00","updated_at":"2026-07-05T09:12:26.561881+00:00"}