{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:VUXXW7KQWINJENFQIBBUECRNUH","short_pith_number":"pith:VUXXW7KQ","schema_version":"1.0","canonical_sha256":"ad2f7b7d50b21a9234b04043420a2da1fe6cf7a8b7a61a6bccb2d6a0cc8f8552","source":{"kind":"arxiv","id":"2407.21424","version":2},"attestation_state":"computed","paper":{"title":"Cost-Effective Hallucination Detection for LLMs","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG","stat.ML"],"primary_cat":"cs.CL","authors_text":"Bryan Wang, Gianluca Detommaso, Giovanni Zappella, Jinmiao Fu, Shaoyuan Xu, Simon Valentin","submitted_at":"2024-07-31T08:19:06Z","abstract_excerpt":"Large language models (LLMs) can be prone to hallucinations - generating unreliable outputs that are unfaithful to their inputs, external facts or internally inconsistent. In this work, we address several challenges for post-hoc hallucination detection in production settings. Our pipeline for hallucination detection entails: first, producing a confidence score representing the likelihood that a generated answer is a hallucination; second, calibrating the score conditional on attributes of the inputs and candidate response; finally, performing detection by thresholding the calibrated score. We "},"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":"2407.21424","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-07-31T08:19:06Z","cross_cats_sorted":["cs.AI","cs.LG","stat.ML"],"title_canon_sha256":"41884ccdb4701c1745de549fe975328b30f40225c698413b8a8e6e336d4e2152","abstract_canon_sha256":"7cc6fa61a2eb345cead658b98a3a5b3fa3cc0a89c304595d499a2508d2466a86"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:53:54.603397Z","signature_b64":"kpI7qYUT+7gIuqoQhpXZLQFi1yq4fLmjqCYu/GqfT/62DURzualaz3IiG6rdFHBLI3XQGMsWXu9eDKxGVrR1Ag==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"ad2f7b7d50b21a9234b04043420a2da1fe6cf7a8b7a61a6bccb2d6a0cc8f8552","last_reissued_at":"2026-07-05T08:53:54.602861Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:53:54.602861Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Cost-Effective Hallucination Detection for LLMs","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG","stat.ML"],"primary_cat":"cs.CL","authors_text":"Bryan Wang, Gianluca Detommaso, Giovanni Zappella, Jinmiao Fu, Shaoyuan Xu, Simon Valentin","submitted_at":"2024-07-31T08:19:06Z","abstract_excerpt":"Large language models (LLMs) can be prone to hallucinations - generating unreliable outputs that are unfaithful to their inputs, external facts or internally inconsistent. In this work, we address several challenges for post-hoc hallucination detection in production settings. Our pipeline for hallucination detection entails: first, producing a confidence score representing the likelihood that a generated answer is a hallucination; second, calibrating the score conditional on attributes of the inputs and candidate response; finally, performing detection by thresholding the calibrated score. We "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2407.21424","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/2407.21424/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":"2407.21424","created_at":"2026-07-05T08:53:54.602931+00:00"},{"alias_kind":"arxiv_version","alias_value":"2407.21424v2","created_at":"2026-07-05T08:53:54.602931+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2407.21424","created_at":"2026-07-05T08:53:54.602931+00:00"},{"alias_kind":"pith_short_12","alias_value":"VUXXW7KQWINJ","created_at":"2026-07-05T08:53:54.602931+00:00"},{"alias_kind":"pith_short_16","alias_value":"VUXXW7KQWINJENFQ","created_at":"2026-07-05T08:53:54.602931+00:00"},{"alias_kind":"pith_short_8","alias_value":"VUXXW7KQ","created_at":"2026-07-05T08:53:54.602931+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2507.23221","citing_title":"A Single Direction of Truth: An Observer Model's Linear Residual Probe Exposes and Steers Contextual Hallucinations","ref_index":34,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/VUXXW7KQWINJENFQIBBUECRNUH","json":"https://pith.science/pith/VUXXW7KQWINJENFQIBBUECRNUH.json","graph_json":"https://pith.science/api/pith-number/VUXXW7KQWINJENFQIBBUECRNUH/graph.json","events_json":"https://pith.science/api/pith-number/VUXXW7KQWINJENFQIBBUECRNUH/events.json","paper":"https://pith.science/paper/VUXXW7KQ"},"agent_actions":{"view_html":"https://pith.science/pith/VUXXW7KQWINJENFQIBBUECRNUH","download_json":"https://pith.science/pith/VUXXW7KQWINJENFQIBBUECRNUH.json","view_paper":"https://pith.science/paper/VUXXW7KQ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2407.21424&json=true","fetch_graph":"https://pith.science/api/pith-number/VUXXW7KQWINJENFQIBBUECRNUH/graph.json","fetch_events":"https://pith.science/api/pith-number/VUXXW7KQWINJENFQIBBUECRNUH/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/VUXXW7KQWINJENFQIBBUECRNUH/action/timestamp_anchor","attest_storage":"https://pith.science/pith/VUXXW7KQWINJENFQIBBUECRNUH/action/storage_attestation","attest_author":"https://pith.science/pith/VUXXW7KQWINJENFQIBBUECRNUH/action/author_attestation","sign_citation":"https://pith.science/pith/VUXXW7KQWINJENFQIBBUECRNUH/action/citation_signature","submit_replication":"https://pith.science/pith/VUXXW7KQWINJENFQIBBUECRNUH/action/replication_record"}},"created_at":"2026-07-05T08:53:54.602931+00:00","updated_at":"2026-07-05T08:53:54.602931+00:00"}