{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:L5OAMSL2XXCW6VOQBBBU2LFSKH","short_pith_number":"pith:L5OAMSL2","schema_version":"1.0","canonical_sha256":"5f5c06497abdc56f55d008434d2cb251e1171fdf95b263608013e301f820d033","source":{"kind":"arxiv","id":"2502.12414","version":2},"attestation_state":"computed","paper":{"title":"Lost in Transcription, Found in Distribution Shift: Demystifying Hallucination in Speech Foundation Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Abdul Waheed, Bhiksha Raj, Hanin Atwany, Monojit Choudhury, Rita Singh","submitted_at":"2025-02-18T01:25:39Z","abstract_excerpt":"Speech foundation models trained at a massive scale, both in terms of model and data size, result in robust systems capable of performing multiple speech tasks, including automatic speech recognition (ASR). These models transcend language and domain barriers, yet effectively measuring their performance remains a challenge. Traditional metrics like word error rate (WER) and character error rate (CER) are commonly used to evaluate ASR performance but often fail to reflect transcription quality in critical contexts, particularly when detecting fabricated outputs. This phenomenon, known as halluci"},"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":"2502.12414","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-02-18T01:25:39Z","cross_cats_sorted":[],"title_canon_sha256":"0c8499221c08e15a30ec0fa58d4e3aa3ba9dba995bda17d67d2fc91e4947bf49","abstract_canon_sha256":"ed2eee49b1be5574c4d25bbeb4d563d1432bf0a3ae80aba4e52587d8df7af3d1"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:16:06.301531Z","signature_b64":"zYcqndPWlKr+UHSZxcakAyaKQPgN39C6pqZ+TSQBp8yfbc/AJA1bED3KSq4aK8F0REJbZVdrwDqCcKkykRxoAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"5f5c06497abdc56f55d008434d2cb251e1171fdf95b263608013e301f820d033","last_reissued_at":"2026-07-05T11:16:06.300977Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:16:06.300977Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Lost in Transcription, Found in Distribution Shift: Demystifying Hallucination in Speech Foundation Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Abdul Waheed, Bhiksha Raj, Hanin Atwany, Monojit Choudhury, Rita Singh","submitted_at":"2025-02-18T01:25:39Z","abstract_excerpt":"Speech foundation models trained at a massive scale, both in terms of model and data size, result in robust systems capable of performing multiple speech tasks, including automatic speech recognition (ASR). These models transcend language and domain barriers, yet effectively measuring their performance remains a challenge. Traditional metrics like word error rate (WER) and character error rate (CER) are commonly used to evaluate ASR performance but often fail to reflect transcription quality in critical contexts, particularly when detecting fabricated outputs. This phenomenon, known as halluci"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2502.12414","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/2502.12414/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":"2502.12414","created_at":"2026-07-05T11:16:06.301046+00:00"},{"alias_kind":"arxiv_version","alias_value":"2502.12414v2","created_at":"2026-07-05T11:16:06.301046+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2502.12414","created_at":"2026-07-05T11:16:06.301046+00:00"},{"alias_kind":"pith_short_12","alias_value":"L5OAMSL2XXCW","created_at":"2026-07-05T11:16:06.301046+00:00"},{"alias_kind":"pith_short_16","alias_value":"L5OAMSL2XXCW6VOQ","created_at":"2026-07-05T11:16:06.301046+00:00"},{"alias_kind":"pith_short_8","alias_value":"L5OAMSL2","created_at":"2026-07-05T11:16:06.301046+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":4,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2607.05364","citing_title":"REDDIT: Correcting Model-Generated Timestamp Drift in ASR without Forgetting via Replay-Based Distribution Editing","ref_index":22,"is_internal_anchor":true},{"citing_arxiv_id":"2606.23060","citing_title":"From Text Metrics to Model Internals: A Study of Whisper ASR Hallucination Detection","ref_index":24,"is_internal_anchor":false},{"citing_arxiv_id":"2606.23048","citing_title":"HALAS: A Human-Annotated Dataset of Hallucinations of Modern ASR Systems","ref_index":13,"is_internal_anchor":false},{"citing_arxiv_id":"2604.08591","citing_title":"From Dispersion to Attraction: Spectral Dynamics of Hallucination Across Whisper Model Scales","ref_index":26,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/L5OAMSL2XXCW6VOQBBBU2LFSKH","json":"https://pith.science/pith/L5OAMSL2XXCW6VOQBBBU2LFSKH.json","graph_json":"https://pith.science/api/pith-number/L5OAMSL2XXCW6VOQBBBU2LFSKH/graph.json","events_json":"https://pith.science/api/pith-number/L5OAMSL2XXCW6VOQBBBU2LFSKH/events.json","paper":"https://pith.science/paper/L5OAMSL2"},"agent_actions":{"view_html":"https://pith.science/pith/L5OAMSL2XXCW6VOQBBBU2LFSKH","download_json":"https://pith.science/pith/L5OAMSL2XXCW6VOQBBBU2LFSKH.json","view_paper":"https://pith.science/paper/L5OAMSL2","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2502.12414&json=true","fetch_graph":"https://pith.science/api/pith-number/L5OAMSL2XXCW6VOQBBBU2LFSKH/graph.json","fetch_events":"https://pith.science/api/pith-number/L5OAMSL2XXCW6VOQBBBU2LFSKH/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/L5OAMSL2XXCW6VOQBBBU2LFSKH/action/timestamp_anchor","attest_storage":"https://pith.science/pith/L5OAMSL2XXCW6VOQBBBU2LFSKH/action/storage_attestation","attest_author":"https://pith.science/pith/L5OAMSL2XXCW6VOQBBBU2LFSKH/action/author_attestation","sign_citation":"https://pith.science/pith/L5OAMSL2XXCW6VOQBBBU2LFSKH/action/citation_signature","submit_replication":"https://pith.science/pith/L5OAMSL2XXCW6VOQBBBU2LFSKH/action/replication_record"}},"created_at":"2026-07-05T11:16:06.301046+00:00","updated_at":"2026-07-05T11:16:06.301046+00:00"}