{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:3PUGFFXKO5TNAKOP2DTJZ2WEBC","short_pith_number":"pith:3PUGFFXK","schema_version":"1.0","canonical_sha256":"dbe86296ea7766d029cfd0e69ceac4089e43c0c981e753ba3171fe6e8cf805b9","source":{"kind":"arxiv","id":"2310.14971","version":1},"attestation_state":"computed","paper":{"title":"Penalty Decoding: Well Suppress the Self-Reinforcement Effect in Open-Ended Text Generation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Hongkun Hao, Rui Wang, Wenhong Zhu","submitted_at":"2023-10-23T14:20:04Z","abstract_excerpt":"The decoding algorithm is critical for open-ended text generation, transforming latent representations into coherent and meaningful outputs. This paper investigates the self-reinforcement effect in text generation and the effectiveness of a repetition penalty to mitigate it. However, determining the optimal repetition penalty value is challenging. To tackle this, we propose a forgetting mechanism that disregards distant tokens, reducing the burden of penalty selection. In addition, we introduce a length penalty to address overly short sentences caused by excessive penalties. Our penalty decodi"},"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":"2310.14971","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-10-23T14:20:04Z","cross_cats_sorted":[],"title_canon_sha256":"88000d3655cc3df37089ce907f8304d27b75d9ad333918ba7c6086b0a7807d8e","abstract_canon_sha256":"f9a1e294f82cbbbebf0067036f939f508d00f85420200351701dc0976400cfba"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:04:04.339692Z","signature_b64":"6sF/4+Mz7OYI1vJaXNmjzbQeLdheecWpwQP7y1xYzQ2BbblmuEEFes4mAgphANkoNFdc+W0wHL1sWwhLbt1mCg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"dbe86296ea7766d029cfd0e69ceac4089e43c0c981e753ba3171fe6e8cf805b9","last_reissued_at":"2026-07-05T07:04:04.339167Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:04:04.339167Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Penalty Decoding: Well Suppress the Self-Reinforcement Effect in Open-Ended Text Generation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Hongkun Hao, Rui Wang, Wenhong Zhu","submitted_at":"2023-10-23T14:20:04Z","abstract_excerpt":"The decoding algorithm is critical for open-ended text generation, transforming latent representations into coherent and meaningful outputs. This paper investigates the self-reinforcement effect in text generation and the effectiveness of a repetition penalty to mitigate it. However, determining the optimal repetition penalty value is challenging. To tackle this, we propose a forgetting mechanism that disregards distant tokens, reducing the burden of penalty selection. In addition, we introduce a length penalty to address overly short sentences caused by excessive penalties. Our penalty decodi"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2310.14971","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/2310.14971/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":"2310.14971","created_at":"2026-07-05T07:04:04.339232+00:00"},{"alias_kind":"arxiv_version","alias_value":"2310.14971v1","created_at":"2026-07-05T07:04:04.339232+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2310.14971","created_at":"2026-07-05T07:04:04.339232+00:00"},{"alias_kind":"pith_short_12","alias_value":"3PUGFFXKO5TN","created_at":"2026-07-05T07:04:04.339232+00:00"},{"alias_kind":"pith_short_16","alias_value":"3PUGFFXKO5TNAKOP","created_at":"2026-07-05T07:04:04.339232+00:00"},{"alias_kind":"pith_short_8","alias_value":"3PUGFFXK","created_at":"2026-07-05T07:04:04.339232+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.11922","citing_title":"StepCodeReasoner: Aligning Code Reasoning with Stepwise Execution Traces via Reinforcement Learning","ref_index":17,"is_internal_anchor":false},{"citing_arxiv_id":"2604.17323","citing_title":"A Universal Avoidance Method for Diverse Multi-branch Generation","ref_index":30,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/3PUGFFXKO5TNAKOP2DTJZ2WEBC","json":"https://pith.science/pith/3PUGFFXKO5TNAKOP2DTJZ2WEBC.json","graph_json":"https://pith.science/api/pith-number/3PUGFFXKO5TNAKOP2DTJZ2WEBC/graph.json","events_json":"https://pith.science/api/pith-number/3PUGFFXKO5TNAKOP2DTJZ2WEBC/events.json","paper":"https://pith.science/paper/3PUGFFXK"},"agent_actions":{"view_html":"https://pith.science/pith/3PUGFFXKO5TNAKOP2DTJZ2WEBC","download_json":"https://pith.science/pith/3PUGFFXKO5TNAKOP2DTJZ2WEBC.json","view_paper":"https://pith.science/paper/3PUGFFXK","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2310.14971&json=true","fetch_graph":"https://pith.science/api/pith-number/3PUGFFXKO5TNAKOP2DTJZ2WEBC/graph.json","fetch_events":"https://pith.science/api/pith-number/3PUGFFXKO5TNAKOP2DTJZ2WEBC/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/3PUGFFXKO5TNAKOP2DTJZ2WEBC/action/timestamp_anchor","attest_storage":"https://pith.science/pith/3PUGFFXKO5TNAKOP2DTJZ2WEBC/action/storage_attestation","attest_author":"https://pith.science/pith/3PUGFFXKO5TNAKOP2DTJZ2WEBC/action/author_attestation","sign_citation":"https://pith.science/pith/3PUGFFXKO5TNAKOP2DTJZ2WEBC/action/citation_signature","submit_replication":"https://pith.science/pith/3PUGFFXKO5TNAKOP2DTJZ2WEBC/action/replication_record"}},"created_at":"2026-07-05T07:04:04.339232+00:00","updated_at":"2026-07-05T07:04:04.339232+00:00"}