{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:YH3GOGQEA6JECXZSX2FNRYZMOT","short_pith_number":"pith:YH3GOGQE","schema_version":"1.0","canonical_sha256":"c1f6671a040792415f32be8ad8e32c74e800711d49eb0cd4f6e88abc629afc48","source":{"kind":"arxiv","id":"2509.02170","version":2},"attestation_state":"computed","paper":{"title":"Avoidance Decoding for Diverse Multi-Branch Story Generation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Kyeongman Park, Kyomin Jung, Nakyeong Yang","submitted_at":"2025-09-02T10:22:46Z","abstract_excerpt":"Large Language Models (LLMs) often generate repetitive and monotonous outputs, especially in tasks like story generation, due to limited creative diversity when given the same input prompt. To address this challenge, we propose a novel decoding strategy, Avoidance Decoding, that modifies token logits by penalizing similarity to previously generated outputs, thereby encouraging more diverse multi-branch stories. This penalty adaptively balances two similarity measures: (1) Concept-level Similarity Penalty, which is prioritized in early stages to diversify initial story concepts, and (2) Narrati"},"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":"2509.02170","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-09-02T10:22:46Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"995a886861fa01f228a288ac9ca945fef4e28f8ef243d5eea92a12a7f3a92ce9","abstract_canon_sha256":"79c998da2f71829c7d4b547f855327ea3d1b4444570a7f46e1d5dfe6db1b1f97"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T12:03:53.053745Z","signature_b64":"jzx+OSKEp8XPZeFOmXSgs+rRntnHTsr3OrdUuDR/4C28XJCKq4UaPFGPk8yPhghmIjBxJYdZ8JCuMBFoeliXCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"c1f6671a040792415f32be8ad8e32c74e800711d49eb0cd4f6e88abc629afc48","last_reissued_at":"2026-07-05T12:03:53.053230Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T12:03:53.053230Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Avoidance Decoding for Diverse Multi-Branch Story Generation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Kyeongman Park, Kyomin Jung, Nakyeong Yang","submitted_at":"2025-09-02T10:22:46Z","abstract_excerpt":"Large Language Models (LLMs) often generate repetitive and monotonous outputs, especially in tasks like story generation, due to limited creative diversity when given the same input prompt. To address this challenge, we propose a novel decoding strategy, Avoidance Decoding, that modifies token logits by penalizing similarity to previously generated outputs, thereby encouraging more diverse multi-branch stories. This penalty adaptively balances two similarity measures: (1) Concept-level Similarity Penalty, which is prioritized in early stages to diversify initial story concepts, and (2) Narrati"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2509.02170","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/2509.02170/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":"2509.02170","created_at":"2026-07-05T12:03:53.053291+00:00"},{"alias_kind":"arxiv_version","alias_value":"2509.02170v2","created_at":"2026-07-05T12:03:53.053291+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2509.02170","created_at":"2026-07-05T12:03:53.053291+00:00"},{"alias_kind":"pith_short_12","alias_value":"YH3GOGQEA6JE","created_at":"2026-07-05T12:03:53.053291+00:00"},{"alias_kind":"pith_short_16","alias_value":"YH3GOGQEA6JECXZS","created_at":"2026-07-05T12:03:53.053291+00:00"},{"alias_kind":"pith_short_8","alias_value":"YH3GOGQE","created_at":"2026-07-05T12:03:53.053291+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2604.17323","citing_title":"A Universal Avoidance Method for Diverse Multi-branch Generation","ref_index":19,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/YH3GOGQEA6JECXZSX2FNRYZMOT","json":"https://pith.science/pith/YH3GOGQEA6JECXZSX2FNRYZMOT.json","graph_json":"https://pith.science/api/pith-number/YH3GOGQEA6JECXZSX2FNRYZMOT/graph.json","events_json":"https://pith.science/api/pith-number/YH3GOGQEA6JECXZSX2FNRYZMOT/events.json","paper":"https://pith.science/paper/YH3GOGQE"},"agent_actions":{"view_html":"https://pith.science/pith/YH3GOGQEA6JECXZSX2FNRYZMOT","download_json":"https://pith.science/pith/YH3GOGQEA6JECXZSX2FNRYZMOT.json","view_paper":"https://pith.science/paper/YH3GOGQE","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2509.02170&json=true","fetch_graph":"https://pith.science/api/pith-number/YH3GOGQEA6JECXZSX2FNRYZMOT/graph.json","fetch_events":"https://pith.science/api/pith-number/YH3GOGQEA6JECXZSX2FNRYZMOT/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/YH3GOGQEA6JECXZSX2FNRYZMOT/action/timestamp_anchor","attest_storage":"https://pith.science/pith/YH3GOGQEA6JECXZSX2FNRYZMOT/action/storage_attestation","attest_author":"https://pith.science/pith/YH3GOGQEA6JECXZSX2FNRYZMOT/action/author_attestation","sign_citation":"https://pith.science/pith/YH3GOGQEA6JECXZSX2FNRYZMOT/action/citation_signature","submit_replication":"https://pith.science/pith/YH3GOGQEA6JECXZSX2FNRYZMOT/action/replication_record"}},"created_at":"2026-07-05T12:03:53.053291+00:00","updated_at":"2026-07-05T12:03:53.053291+00:00"}