{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:KAS6LSWDF734IK46MFVIILTV7B","short_pith_number":"pith:KAS6LSWD","schema_version":"1.0","canonical_sha256":"5025e5cac32ff7c42b9e616a842e75f869be09ca4a9ab52803852af3ef87c16b","source":{"kind":"arxiv","id":"2204.05424","version":3},"attestation_state":"computed","paper":{"title":"A Call for Clarity in Beam Search: How It Works and When It Stops","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Dragomir Radev, Jungo Kasai, Keisuke Sakaguchi, Noah A. Smith, Ronan Le Bras, Yejin Choi","submitted_at":"2022-04-11T22:03:44Z","abstract_excerpt":"Text generation with beam search has proven successful in a wide range of applications. We point out that, though largely overlooked in the literature, the commonly-used implementation of beam decoding (e.g., Hugging Face Transformers and fairseq) uses a first come, first served heuristic: it keeps a set of already completed sequences over time steps and stops when the size of this set reaches the beam size. Based on this finding, we introduce a patience factor, a simple modification to this beam decoding implementation, that generalizes the stopping criterion and provides flexibility to the d"},"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":"2204.05424","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2022-04-11T22:03:44Z","cross_cats_sorted":[],"title_canon_sha256":"aa54deedc3a0f25e425d7b18dff33d019e8a6e05c32a5e12dfa846892e25497d","abstract_canon_sha256":"37f3e253749adba863578fcf91265135c45b35927072e8aba42b3a57fe0d5070"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:49:58.367311Z","signature_b64":"/YWXu7rIbbxCarWPUW/32CLWyOXrXOmChPcYakbIWk9qlLnbJwz9f9eZbBtQvufNAU5RmcoX0+0vFn7/n4PmCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"5025e5cac32ff7c42b9e616a842e75f869be09ca4a9ab52803852af3ef87c16b","last_reissued_at":"2026-07-05T07:49:58.366719Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:49:58.366719Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"A Call for Clarity in Beam Search: How It Works and When It Stops","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Dragomir Radev, Jungo Kasai, Keisuke Sakaguchi, Noah A. Smith, Ronan Le Bras, Yejin Choi","submitted_at":"2022-04-11T22:03:44Z","abstract_excerpt":"Text generation with beam search has proven successful in a wide range of applications. We point out that, though largely overlooked in the literature, the commonly-used implementation of beam decoding (e.g., Hugging Face Transformers and fairseq) uses a first come, first served heuristic: it keeps a set of already completed sequences over time steps and stops when the size of this set reaches the beam size. Based on this finding, we introduce a patience factor, a simple modification to this beam decoding implementation, that generalizes the stopping criterion and provides flexibility to the d"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2204.05424","kind":"arxiv","version":3},"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/2204.05424/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":"2204.05424","created_at":"2026-07-05T07:49:58.366783+00:00"},{"alias_kind":"arxiv_version","alias_value":"2204.05424v3","created_at":"2026-07-05T07:49:58.366783+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2204.05424","created_at":"2026-07-05T07:49:58.366783+00:00"},{"alias_kind":"pith_short_12","alias_value":"KAS6LSWDF734","created_at":"2026-07-05T07:49:58.366783+00:00"},{"alias_kind":"pith_short_16","alias_value":"KAS6LSWDF734IK46","created_at":"2026-07-05T07:49:58.366783+00:00"},{"alias_kind":"pith_short_8","alias_value":"KAS6LSWD","created_at":"2026-07-05T07:49:58.366783+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2506.11091","citing_title":"Customizing Speech Recognition Model with Large Language Model Feedback","ref_index":36,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/KAS6LSWDF734IK46MFVIILTV7B","json":"https://pith.science/pith/KAS6LSWDF734IK46MFVIILTV7B.json","graph_json":"https://pith.science/api/pith-number/KAS6LSWDF734IK46MFVIILTV7B/graph.json","events_json":"https://pith.science/api/pith-number/KAS6LSWDF734IK46MFVIILTV7B/events.json","paper":"https://pith.science/paper/KAS6LSWD"},"agent_actions":{"view_html":"https://pith.science/pith/KAS6LSWDF734IK46MFVIILTV7B","download_json":"https://pith.science/pith/KAS6LSWDF734IK46MFVIILTV7B.json","view_paper":"https://pith.science/paper/KAS6LSWD","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2204.05424&json=true","fetch_graph":"https://pith.science/api/pith-number/KAS6LSWDF734IK46MFVIILTV7B/graph.json","fetch_events":"https://pith.science/api/pith-number/KAS6LSWDF734IK46MFVIILTV7B/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/KAS6LSWDF734IK46MFVIILTV7B/action/timestamp_anchor","attest_storage":"https://pith.science/pith/KAS6LSWDF734IK46MFVIILTV7B/action/storage_attestation","attest_author":"https://pith.science/pith/KAS6LSWDF734IK46MFVIILTV7B/action/author_attestation","sign_citation":"https://pith.science/pith/KAS6LSWDF734IK46MFVIILTV7B/action/citation_signature","submit_replication":"https://pith.science/pith/KAS6LSWDF734IK46MFVIILTV7B/action/replication_record"}},"created_at":"2026-07-05T07:49:58.366783+00:00","updated_at":"2026-07-05T07:49:58.366783+00:00"}