{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:3IE37OUWYPQSMRHFCUPUEWNUNH","short_pith_number":"pith:3IE37OUW","schema_version":"1.0","canonical_sha256":"da09bfba96c3e12644e5151f4259b469f1c21782d961719cb7da537a89e82b24","source":{"kind":"arxiv","id":"2605.10930","version":2},"attestation_state":"computed","paper":{"title":"Evaluating the False Trust Engendered by LLM Explanations","license":"http://creativecommons.org/licenses/by/4.0/","headline":"Reasoning traces and post-hoc explanations from LLMs increase user acceptance of answers whether correct or incorrect, while only dual explanations improve users' ability to tell the difference.","cross_cats":[],"primary_cat":"cs.HC","authors_text":"Subbarao Kambhampati, Upasana Biswas, Vardhan Palod","submitted_at":"2026-05-11T17:58:12Z","abstract_excerpt":"Large Language Models (LLMs) and Large Reasoning Models (LRMs) are increasingly used for critical tasks, yet they provide no guarantees about the correctness of their solutions. Users must decide whether to trust the model's answer, aided by reasoning traces, their summaries, or post-hoc generated explanations. These reasoning traces, despite evidence that they are neither faithful representations of the model's computations nor necessarily semantically meaningful, are often interpreted as provenance explanations. It is unclear whether explanations or reasoning traces help users identify when "},"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":true},"canonical_record":{"source":{"id":"2605.10930","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.HC","submitted_at":"2026-05-11T17:58:12Z","cross_cats_sorted":[],"title_canon_sha256":"c5626129d5149c04c92db68bbdb213b16d502872d5a74a69ce93174d5183fc4b","abstract_canon_sha256":"f327e1a5cd50f6f794e5f5a33856736c2448fde57f4076fd38948968298636e0"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-20T00:02:13.022487Z","signature_b64":"XGuIWZ1e48qHyir8cH2SiOgMGwIzWb+XKTH5L8fSl0Mne4Ntcy+b9jkj6qwgtCw4TF3epo2w5t5655wenp7kDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"da09bfba96c3e12644e5151f4259b469f1c21782d961719cb7da537a89e82b24","last_reissued_at":"2026-05-20T00:02:13.021655Z","signature_status":"signed_v1","first_computed_at":"2026-05-20T00:02:13.021655Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Evaluating the False Trust Engendered by LLM Explanations","license":"http://creativecommons.org/licenses/by/4.0/","headline":"Reasoning traces and post-hoc explanations from LLMs increase user acceptance of answers whether correct or incorrect, while only dual explanations improve users' ability to tell the difference.","cross_cats":[],"primary_cat":"cs.HC","authors_text":"Subbarao Kambhampati, Upasana Biswas, Vardhan Palod","submitted_at":"2026-05-11T17:58:12Z","abstract_excerpt":"Large Language Models (LLMs) and Large Reasoning Models (LRMs) are increasingly used for critical tasks, yet they provide no guarantees about the correctness of their solutions. Users must decide whether to trust the model's answer, aided by reasoning traces, their summaries, or post-hoc generated explanations. These reasoning traces, despite evidence that they are neither faithful representations of the model's computations nor necessarily semantically meaningful, are often interpreted as provenance explanations. It is unclear whether explanations or reasoning traces help users identify when "},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"reasoning traces and post-hoc explanations are persuasive but not informative: they increase user acceptance of LLM predictions regardless of their correctness. In contrast, dual explanation is the only condition that genuinely improves users' ability to distinguish correct from incorrect AI outputs.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"The assumption that the simulated setting where users do not have the means to verify the solution and the between-subject design with chosen tasks accurately measures real-world false trust and generalizes to critical task scenarios.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"A user study finds that LLM reasoning traces and post-hoc explanations create false trust by increasing acceptance of incorrect answers, whereas contrastive dual explanations improve users' ability to detect errors.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Reasoning traces and post-hoc explanations from LLMs increase user acceptance of answers whether correct or incorrect, while only dual explanations improve users' ability to tell the difference.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"e44282541e42c2f6cc23dc18b9de1bf17cfdfb6a3319d2f73dca3e6cec6ad193"},"source":{"id":"2605.10930","kind":"arxiv","version":2},"verdict":{"id":"5ae667ca-c289-4265-bfe5-06b9d217b50e","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-12T03:18:28.568972Z","strongest_claim":"reasoning traces and post-hoc explanations are persuasive but not informative: they increase user acceptance of LLM predictions regardless of their correctness. In contrast, dual explanation is the only condition that genuinely improves users' ability to distinguish correct from incorrect AI outputs.","one_line_summary":"A user study finds that LLM reasoning traces and post-hoc explanations create false trust by increasing acceptance of incorrect answers, whereas contrastive dual explanations improve users' ability to detect errors.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"The assumption that the simulated setting where users do not have the means to verify the solution and the between-subject design with chosen tasks accurately measures real-world false trust and generalizes to critical task scenarios.","pith_extraction_headline":"Reasoning traces and post-hoc explanations from LLMs increase user acceptance of answers whether correct or incorrect, while only dual explanations improve users' ability to tell the difference."},"integrity":{"clean":false,"summary":{"advisory":1,"critical":0,"by_detector":{"doi_compliance":{"total":1,"advisory":1,"critical":0,"informational":0}},"informational":0},"endpoint":"/pith/2605.10930/integrity.json","findings":[{"note":"DOI in the printed bibliography is fragmented by whitespace or line breaks. A longer candidate (10.18653/v1/2024.naacl-long.81.URLhttps://aclanthology.org/2024.naacl-long.81/) was visible in the surrounding text but could not be confirmed against doi.org as printed.","detector":"doi_compliance","severity":"advisory","ref_index":24,"audited_at":"2026-05-19T08:53:51.085219Z","detected_doi":"10.18653/v1/2024.naacl-long.81.URLhttps://aclanthology.org/2024.naacl-long.81/","finding_type":"recoverable_identifier","verdict_class":"incontrovertible","detected_arxiv_id":null}],"available":true,"detectors_run":[{"name":"ai_meta_artifact","ran_at":"2026-05-19T13:36:58.657810Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_title_agreement","ran_at":"2026-05-19T10:31:17.157381Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_compliance","ran_at":"2026-05-19T08:53:51.085219Z","status":"completed","version":"1.0.0","findings_count":1}],"snapshot_sha256":"c3dee2ab14b863397e982b4374548834d30745848e35d8072a2e55582b225563"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":2,"snapshot_sha256":"c4946d7ae42820a65b900c3b62b1ac03862972f1a3de6ba447e9528e1d46a0f8"},"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":"2605.10930","created_at":"2026-05-20T00:02:13.021800+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.10930v2","created_at":"2026-05-20T00:02:13.021800+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.10930","created_at":"2026-05-20T00:02:13.021800+00:00"},{"alias_kind":"pith_short_12","alias_value":"3IE37OUWYPQS","created_at":"2026-05-20T00:02:13.021800+00:00"},{"alias_kind":"pith_short_16","alias_value":"3IE37OUWYPQSMRHF","created_at":"2026-05-20T00:02:13.021800+00:00"},{"alias_kind":"pith_short_8","alias_value":"3IE37OUW","created_at":"2026-05-20T00:02:13.021800+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":2,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/3IE37OUWYPQSMRHFCUPUEWNUNH","json":"https://pith.science/pith/3IE37OUWYPQSMRHFCUPUEWNUNH.json","graph_json":"https://pith.science/api/pith-number/3IE37OUWYPQSMRHFCUPUEWNUNH/graph.json","events_json":"https://pith.science/api/pith-number/3IE37OUWYPQSMRHFCUPUEWNUNH/events.json","paper":"https://pith.science/paper/3IE37OUW"},"agent_actions":{"view_html":"https://pith.science/pith/3IE37OUWYPQSMRHFCUPUEWNUNH","download_json":"https://pith.science/pith/3IE37OUWYPQSMRHFCUPUEWNUNH.json","view_paper":"https://pith.science/paper/3IE37OUW","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.10930&json=true","fetch_graph":"https://pith.science/api/pith-number/3IE37OUWYPQSMRHFCUPUEWNUNH/graph.json","fetch_events":"https://pith.science/api/pith-number/3IE37OUWYPQSMRHFCUPUEWNUNH/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/3IE37OUWYPQSMRHFCUPUEWNUNH/action/timestamp_anchor","attest_storage":"https://pith.science/pith/3IE37OUWYPQSMRHFCUPUEWNUNH/action/storage_attestation","attest_author":"https://pith.science/pith/3IE37OUWYPQSMRHFCUPUEWNUNH/action/author_attestation","sign_citation":"https://pith.science/pith/3IE37OUWYPQSMRHFCUPUEWNUNH/action/citation_signature","submit_replication":"https://pith.science/pith/3IE37OUWYPQSMRHFCUPUEWNUNH/action/replication_record"}},"created_at":"2026-05-20T00:02:13.021800+00:00","updated_at":"2026-05-20T00:02:13.021800+00:00"}