{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:NNWVVLLO3YY42AYOX2LGB6NTXR","short_pith_number":"pith:NNWVVLLO","schema_version":"1.0","canonical_sha256":"6b6d5aad6ede31cd030ebe9660f9b3bc5676932f985788a72f97772b011eeaae","source":{"kind":"arxiv","id":"2509.10004","version":1},"attestation_state":"computed","paper":{"title":"Unsupervised Hallucination Detection by Inspecting Reasoning Processes","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Anh Tuan Luu, Ponhvoan Srey, Xiaobao Wu","submitted_at":"2025-09-12T06:58:17Z","abstract_excerpt":"Unsupervised hallucination detection aims to identify hallucinated content generated by large language models (LLMs) without relying on labeled data. While unsupervised methods have gained popularity by eliminating labor-intensive human annotations, they frequently rely on proxy signals unrelated to factual correctness. This misalignment biases detection probes toward superficial or non-truth-related aspects, limiting generalizability across datasets and scenarios. To overcome these limitations, we propose IRIS, an unsupervised hallucination detection framework, leveraging internal representat"},"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":"2509.10004","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-09-12T06:58:17Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"c09b0780ca5471d7f09e7112a5c4355d8ff6bfa1bf685cfcc045a2dc982b36e9","abstract_canon_sha256":"b57b39d531649f37dd4d46739b134d6fd5e48681c17489f22757f0d619170001"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T12:09:53.544519Z","signature_b64":"/ue11P9+L24LvESdYo0EM004qFZRz9x/GpGyy8Ii9JxrujaagUnvJbud1jaUttL2dX7ReCjBwtrziSQ3USvWCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"6b6d5aad6ede31cd030ebe9660f9b3bc5676932f985788a72f97772b011eeaae","last_reissued_at":"2026-07-05T12:09:53.544090Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T12:09:53.544090Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Unsupervised Hallucination Detection by Inspecting Reasoning Processes","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Anh Tuan Luu, Ponhvoan Srey, Xiaobao Wu","submitted_at":"2025-09-12T06:58:17Z","abstract_excerpt":"Unsupervised hallucination detection aims to identify hallucinated content generated by large language models (LLMs) without relying on labeled data. While unsupervised methods have gained popularity by eliminating labor-intensive human annotations, they frequently rely on proxy signals unrelated to factual correctness. This misalignment biases detection probes toward superficial or non-truth-related aspects, limiting generalizability across datasets and scenarios. To overcome these limitations, we propose IRIS, an unsupervised hallucination detection framework, leveraging internal representat"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2509.10004","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/2509.10004/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":"2509.10004","created_at":"2026-07-05T12:09:53.544148+00:00"},{"alias_kind":"arxiv_version","alias_value":"2509.10004v1","created_at":"2026-07-05T12:09:53.544148+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2509.10004","created_at":"2026-07-05T12:09:53.544148+00:00"},{"alias_kind":"pith_short_12","alias_value":"NNWVVLLO3YY4","created_at":"2026-07-05T12:09:53.544148+00:00"},{"alias_kind":"pith_short_16","alias_value":"NNWVVLLO3YY42AYO","created_at":"2026-07-05T12:09:53.544148+00:00"},{"alias_kind":"pith_short_8","alias_value":"NNWVVLLO","created_at":"2026-07-05T12:09:53.544148+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/NNWVVLLO3YY42AYOX2LGB6NTXR","json":"https://pith.science/pith/NNWVVLLO3YY42AYOX2LGB6NTXR.json","graph_json":"https://pith.science/api/pith-number/NNWVVLLO3YY42AYOX2LGB6NTXR/graph.json","events_json":"https://pith.science/api/pith-number/NNWVVLLO3YY42AYOX2LGB6NTXR/events.json","paper":"https://pith.science/paper/NNWVVLLO"},"agent_actions":{"view_html":"https://pith.science/pith/NNWVVLLO3YY42AYOX2LGB6NTXR","download_json":"https://pith.science/pith/NNWVVLLO3YY42AYOX2LGB6NTXR.json","view_paper":"https://pith.science/paper/NNWVVLLO","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2509.10004&json=true","fetch_graph":"https://pith.science/api/pith-number/NNWVVLLO3YY42AYOX2LGB6NTXR/graph.json","fetch_events":"https://pith.science/api/pith-number/NNWVVLLO3YY42AYOX2LGB6NTXR/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/NNWVVLLO3YY42AYOX2LGB6NTXR/action/timestamp_anchor","attest_storage":"https://pith.science/pith/NNWVVLLO3YY42AYOX2LGB6NTXR/action/storage_attestation","attest_author":"https://pith.science/pith/NNWVVLLO3YY42AYOX2LGB6NTXR/action/author_attestation","sign_citation":"https://pith.science/pith/NNWVVLLO3YY42AYOX2LGB6NTXR/action/citation_signature","submit_replication":"https://pith.science/pith/NNWVVLLO3YY42AYOX2LGB6NTXR/action/replication_record"}},"created_at":"2026-07-05T12:09:53.544148+00:00","updated_at":"2026-07-05T12:09:53.544148+00:00"}