{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:W7C56FH7XUNN7LPYU2QTJT3Z6I","short_pith_number":"pith:W7C56FH7","schema_version":"1.0","canonical_sha256":"b7c5df14ffbd1adfadf8a6a134cf79f22b3d4e49bb765210bc8b2b821798eb82","source":{"kind":"arxiv","id":"2606.21917","version":1},"attestation_state":"computed","paper":{"title":"Pre-Generation Hallucination Detection in Large Language Models via Soft-Target Attention Probing","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Alexey Zaytsev, Amina Miftakhova","submitted_at":"2026-06-20T07:31:41Z","abstract_excerpt":"Detecting hallucination risk before generation enables abstention, retrieval augmentation, and routing decisions without incurring the cost of decoding.\n  While prior work has shown that such risk can be estimated from a model's internal representations, existing approaches treat this as binary classification over a single decoded output. We instead formulate it as a risk-estimation problem. Under this formulation, we introduce soft-target supervision based on the empirical answer error rate over stochastically sampled outputs - an estimator we prove to be the unique unbiased minimum-variance "},"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":"2606.21917","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-06-20T07:31:41Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"3dbc7ebd3f5ca9f5c518948ad6ac06a5ead3c5894da1bf03c7c3f8df46f98664","abstract_canon_sha256":"a5065e9c5ef91a22e035a63e40f6c0208a9fa505bfdd31c56f47eb0dc0f7e21e"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-23T02:13:02.821577Z","signature_b64":"w54of+SoubMsKAaXqxK92+jQ11IchHrV4oXq8RGMPRSIJISitDvQjTUDHtqRO/WxqkxwOYY2fSVO1Y0olQL/Dw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"b7c5df14ffbd1adfadf8a6a134cf79f22b3d4e49bb765210bc8b2b821798eb82","last_reissued_at":"2026-06-23T02:13:02.821151Z","signature_status":"signed_v1","first_computed_at":"2026-06-23T02:13:02.821151Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Pre-Generation Hallucination Detection in Large Language Models via Soft-Target Attention Probing","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Alexey Zaytsev, Amina Miftakhova","submitted_at":"2026-06-20T07:31:41Z","abstract_excerpt":"Detecting hallucination risk before generation enables abstention, retrieval augmentation, and routing decisions without incurring the cost of decoding.\n  While prior work has shown that such risk can be estimated from a model's internal representations, existing approaches treat this as binary classification over a single decoded output. We instead formulate it as a risk-estimation problem. Under this formulation, we introduce soft-target supervision based on the empirical answer error rate over stochastically sampled outputs - an estimator we prove to be the unique unbiased minimum-variance "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.21917","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/2606.21917/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":"2606.21917","created_at":"2026-06-23T02:13:02.821212+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.21917v1","created_at":"2026-06-23T02:13:02.821212+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.21917","created_at":"2026-06-23T02:13:02.821212+00:00"},{"alias_kind":"pith_short_12","alias_value":"W7C56FH7XUNN","created_at":"2026-06-23T02:13:02.821212+00:00"},{"alias_kind":"pith_short_16","alias_value":"W7C56FH7XUNN7LPY","created_at":"2026-06-23T02:13:02.821212+00:00"},{"alias_kind":"pith_short_8","alias_value":"W7C56FH7","created_at":"2026-06-23T02:13:02.821212+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/W7C56FH7XUNN7LPYU2QTJT3Z6I","json":"https://pith.science/pith/W7C56FH7XUNN7LPYU2QTJT3Z6I.json","graph_json":"https://pith.science/api/pith-number/W7C56FH7XUNN7LPYU2QTJT3Z6I/graph.json","events_json":"https://pith.science/api/pith-number/W7C56FH7XUNN7LPYU2QTJT3Z6I/events.json","paper":"https://pith.science/paper/W7C56FH7"},"agent_actions":{"view_html":"https://pith.science/pith/W7C56FH7XUNN7LPYU2QTJT3Z6I","download_json":"https://pith.science/pith/W7C56FH7XUNN7LPYU2QTJT3Z6I.json","view_paper":"https://pith.science/paper/W7C56FH7","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.21917&json=true","fetch_graph":"https://pith.science/api/pith-number/W7C56FH7XUNN7LPYU2QTJT3Z6I/graph.json","fetch_events":"https://pith.science/api/pith-number/W7C56FH7XUNN7LPYU2QTJT3Z6I/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/W7C56FH7XUNN7LPYU2QTJT3Z6I/action/timestamp_anchor","attest_storage":"https://pith.science/pith/W7C56FH7XUNN7LPYU2QTJT3Z6I/action/storage_attestation","attest_author":"https://pith.science/pith/W7C56FH7XUNN7LPYU2QTJT3Z6I/action/author_attestation","sign_citation":"https://pith.science/pith/W7C56FH7XUNN7LPYU2QTJT3Z6I/action/citation_signature","submit_replication":"https://pith.science/pith/W7C56FH7XUNN7LPYU2QTJT3Z6I/action/replication_record"}},"created_at":"2026-06-23T02:13:02.821212+00:00","updated_at":"2026-06-23T02:13:02.821212+00:00"}