{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:5XDP7QK2RRFFRXUSK4FAGMW27Y","short_pith_number":"pith:5XDP7QK2","schema_version":"1.0","canonical_sha256":"edc6ffc15a8c4a58de92570a0332dafe19dd9c81eb5657322c460bdc8ed629f3","source":{"kind":"arxiv","id":"2402.09733","version":1},"attestation_state":"computed","paper":{"title":"Do LLMs Know about Hallucination? An Empirical Investigation of LLM's Hidden States","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Hanyu Duan, Kar Yan Tam, Yi Yang","submitted_at":"2024-02-15T06:14:55Z","abstract_excerpt":"Large Language Models (LLMs) can make up answers that are not real, and this is known as hallucination. This research aims to see if, how, and to what extent LLMs are aware of hallucination. More specifically, we check whether and how an LLM reacts differently in its hidden states when it answers a question right versus when it hallucinates. To do this, we introduce an experimental framework which allows examining LLM's hidden states in different hallucination situations. Building upon this framework, we conduct a series of experiments with language models in the LLaMA family (Touvron et al., "},"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":"2402.09733","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-02-15T06:14:55Z","cross_cats_sorted":[],"title_canon_sha256":"5ed7b237be8004aadf9485f8b62fd37089e421b64dab20f08b4c607997193a37","abstract_canon_sha256":"dc66b71444276fdbed15f31a7b0584a4fe42e110e6b6709e27a3b25b087c4f14"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:45:36.134911Z","signature_b64":"uv4TaFrIIMXtpyjFa4Qg6xiLCXjGLbNeB5x96LUbA6/LPH79Fgk1BysxqJhnUyBxcXUcq7i0RHEkdJDPY8e5Bw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"edc6ffc15a8c4a58de92570a0332dafe19dd9c81eb5657322c460bdc8ed629f3","last_reissued_at":"2026-07-05T07:45:36.134455Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:45:36.134455Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Do LLMs Know about Hallucination? An Empirical Investigation of LLM's Hidden States","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Hanyu Duan, Kar Yan Tam, Yi Yang","submitted_at":"2024-02-15T06:14:55Z","abstract_excerpt":"Large Language Models (LLMs) can make up answers that are not real, and this is known as hallucination. This research aims to see if, how, and to what extent LLMs are aware of hallucination. More specifically, we check whether and how an LLM reacts differently in its hidden states when it answers a question right versus when it hallucinates. To do this, we introduce an experimental framework which allows examining LLM's hidden states in different hallucination situations. Building upon this framework, we conduct a series of experiments with language models in the LLaMA family (Touvron et al., "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2402.09733","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/2402.09733/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":"2402.09733","created_at":"2026-07-05T07:45:36.134512+00:00"},{"alias_kind":"arxiv_version","alias_value":"2402.09733v1","created_at":"2026-07-05T07:45:36.134512+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.09733","created_at":"2026-07-05T07:45:36.134512+00:00"},{"alias_kind":"pith_short_12","alias_value":"5XDP7QK2RRFF","created_at":"2026-07-05T07:45:36.134512+00:00"},{"alias_kind":"pith_short_16","alias_value":"5XDP7QK2RRFFRXUS","created_at":"2026-07-05T07:45:36.134512+00:00"},{"alias_kind":"pith_short_8","alias_value":"5XDP7QK2","created_at":"2026-07-05T07:45:36.134512+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":4,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.11640","citing_title":"TAROT: Task-Adaptive Refinement of LLM-prior Graphs for Few-shot Tabular Learning","ref_index":12,"is_internal_anchor":false},{"citing_arxiv_id":"2605.28264","citing_title":"Entropy Distribution as a Fingerprint for Hallucinations in Generative Models","ref_index":10,"is_internal_anchor":false},{"citing_arxiv_id":"2504.00446","citing_title":"Exposing the Ghost in the Transformer: Abnormal Detection for Large Language Models via Hidden State Forensics","ref_index":38,"is_internal_anchor":false},{"citing_arxiv_id":"2604.15945","citing_title":"RAGognizer: Hallucination-Aware Fine-Tuning via Detection Head Integration","ref_index":37,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/5XDP7QK2RRFFRXUSK4FAGMW27Y","json":"https://pith.science/pith/5XDP7QK2RRFFRXUSK4FAGMW27Y.json","graph_json":"https://pith.science/api/pith-number/5XDP7QK2RRFFRXUSK4FAGMW27Y/graph.json","events_json":"https://pith.science/api/pith-number/5XDP7QK2RRFFRXUSK4FAGMW27Y/events.json","paper":"https://pith.science/paper/5XDP7QK2"},"agent_actions":{"view_html":"https://pith.science/pith/5XDP7QK2RRFFRXUSK4FAGMW27Y","download_json":"https://pith.science/pith/5XDP7QK2RRFFRXUSK4FAGMW27Y.json","view_paper":"https://pith.science/paper/5XDP7QK2","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2402.09733&json=true","fetch_graph":"https://pith.science/api/pith-number/5XDP7QK2RRFFRXUSK4FAGMW27Y/graph.json","fetch_events":"https://pith.science/api/pith-number/5XDP7QK2RRFFRXUSK4FAGMW27Y/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/5XDP7QK2RRFFRXUSK4FAGMW27Y/action/timestamp_anchor","attest_storage":"https://pith.science/pith/5XDP7QK2RRFFRXUSK4FAGMW27Y/action/storage_attestation","attest_author":"https://pith.science/pith/5XDP7QK2RRFFRXUSK4FAGMW27Y/action/author_attestation","sign_citation":"https://pith.science/pith/5XDP7QK2RRFFRXUSK4FAGMW27Y/action/citation_signature","submit_replication":"https://pith.science/pith/5XDP7QK2RRFFRXUSK4FAGMW27Y/action/replication_record"}},"created_at":"2026-07-05T07:45:36.134512+00:00","updated_at":"2026-07-05T07:45:36.134512+00:00"}