{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:FTSRDTFLFMS7TW7JTVZPDIUCE6","short_pith_number":"pith:FTSRDTFL","schema_version":"1.0","canonical_sha256":"2ce511ccab2b25f9dbe99d72f1a28227b19c1664995bffc2f8b7ab014626c96f","source":{"kind":"arxiv","id":"2512.24562","version":2},"attestation_state":"computed","paper":{"title":"HaluNet: Learning Hallucination Risk from Internal Signals in LLM Question Answering","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Chaodong Tong, Lei Jiang, Qi Zhang, Yanbing Liu, Zhuojun Jiang","submitted_at":"2025-12-31T02:03:10Z","abstract_excerpt":"Large language models (LLMs) achieve strong question answering (QA) performance but can produce fluent answers unsupported by available evidence. Existing hallucination detectors often rely on external verification, repeated sampling, or test-time judge calls, which can be costly for real-time QA. We propose \\textbf{HaluNet}, a lightweight hallucination risk estimator that uses internal signals from one model generation. HaluNet jointly models token likelihood, predictive entropy, and hidden-state information, allowing probabilistic, distributional, and semantic evidence to inform an answer-le"},"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":"2512.24562","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-12-31T02:03:10Z","cross_cats_sorted":[],"title_canon_sha256":"9b29a144b01431095b1bac10e22f92b80a39be5bf18c5bbe3327d0f88851a011","abstract_canon_sha256":"74fe71be3bbf9600f06927c560928f6b56272f29be529767c4e6c4a102fe9c81"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-29T01:05:00.693746Z","signature_b64":"gHSZG5kmqJv7MGBFkQX1niIwc9rC+m4U9rJESFX8ZF0+z/4OJSkFeHlWlrvmZsYBjeYviqfOda1++deRlEojAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"2ce511ccab2b25f9dbe99d72f1a28227b19c1664995bffc2f8b7ab014626c96f","last_reissued_at":"2026-05-29T01:05:00.692832Z","signature_status":"signed_v1","first_computed_at":"2026-05-29T01:05:00.692832Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"HaluNet: Learning Hallucination Risk from Internal Signals in LLM Question Answering","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Chaodong Tong, Lei Jiang, Qi Zhang, Yanbing Liu, Zhuojun Jiang","submitted_at":"2025-12-31T02:03:10Z","abstract_excerpt":"Large language models (LLMs) achieve strong question answering (QA) performance but can produce fluent answers unsupported by available evidence. Existing hallucination detectors often rely on external verification, repeated sampling, or test-time judge calls, which can be costly for real-time QA. We propose \\textbf{HaluNet}, a lightweight hallucination risk estimator that uses internal signals from one model generation. HaluNet jointly models token likelihood, predictive entropy, and hidden-state information, allowing probabilistic, distributional, and semantic evidence to inform an answer-le"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2512.24562","kind":"arxiv","version":2},"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/2512.24562/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":"2512.24562","created_at":"2026-05-29T01:05:00.692957+00:00"},{"alias_kind":"arxiv_version","alias_value":"2512.24562v2","created_at":"2026-05-29T01:05:00.692957+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2512.24562","created_at":"2026-05-29T01:05:00.692957+00:00"},{"alias_kind":"pith_short_12","alias_value":"FTSRDTFLFMS7","created_at":"2026-05-29T01:05:00.692957+00:00"},{"alias_kind":"pith_short_16","alias_value":"FTSRDTFLFMS7TW7J","created_at":"2026-05-29T01:05:00.692957+00:00"},{"alias_kind":"pith_short_8","alias_value":"FTSRDTFL","created_at":"2026-05-29T01:05:00.692957+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2604.11662","citing_title":"Hidden Failures in Robustness: Why Supervised Uncertainty Quantification Needs Better Evaluation","ref_index":7,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/FTSRDTFLFMS7TW7JTVZPDIUCE6","json":"https://pith.science/pith/FTSRDTFLFMS7TW7JTVZPDIUCE6.json","graph_json":"https://pith.science/api/pith-number/FTSRDTFLFMS7TW7JTVZPDIUCE6/graph.json","events_json":"https://pith.science/api/pith-number/FTSRDTFLFMS7TW7JTVZPDIUCE6/events.json","paper":"https://pith.science/paper/FTSRDTFL"},"agent_actions":{"view_html":"https://pith.science/pith/FTSRDTFLFMS7TW7JTVZPDIUCE6","download_json":"https://pith.science/pith/FTSRDTFLFMS7TW7JTVZPDIUCE6.json","view_paper":"https://pith.science/paper/FTSRDTFL","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2512.24562&json=true","fetch_graph":"https://pith.science/api/pith-number/FTSRDTFLFMS7TW7JTVZPDIUCE6/graph.json","fetch_events":"https://pith.science/api/pith-number/FTSRDTFLFMS7TW7JTVZPDIUCE6/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/FTSRDTFLFMS7TW7JTVZPDIUCE6/action/timestamp_anchor","attest_storage":"https://pith.science/pith/FTSRDTFLFMS7TW7JTVZPDIUCE6/action/storage_attestation","attest_author":"https://pith.science/pith/FTSRDTFLFMS7TW7JTVZPDIUCE6/action/author_attestation","sign_citation":"https://pith.science/pith/FTSRDTFLFMS7TW7JTVZPDIUCE6/action/citation_signature","submit_replication":"https://pith.science/pith/FTSRDTFLFMS7TW7JTVZPDIUCE6/action/replication_record"}},"created_at":"2026-05-29T01:05:00.692957+00:00","updated_at":"2026-05-29T01:05:00.692957+00:00"}