{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:QYRSOM4GY5JN2J3HJN2VHMGPHA","short_pith_number":"pith:QYRSOM4G","schema_version":"1.0","canonical_sha256":"8623273386c752dd27674b7553b0cf381f431e43e8503d40a9540b3c2bffac09","source":{"kind":"arxiv","id":"2505.17694","version":2},"attestation_state":"computed","paper":{"title":"CoDec: Prefix-Shared Decoding Kernel for LLMs","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Chao Fang, Chengying Huan, Chen Tian, Guihai Chen, Kun Yang, Mo Zhou, Rong Gu, Rui Ning, Shaobo Ma, Sheng Zhong, Xi Lin, Xue Li, Zhibin Wang, Zhongfeng Wang, Zhonghui Zhang","submitted_at":"2025-05-23T10:03:28Z","abstract_excerpt":"Prefix-sharing among multiple prompts presents opportunities to combine the operations of the shared prefix, while attention computation in the decode stage, which becomes a critical bottleneck with increasing context lengths, is a memory-intensive process requiring heavy memory access on the key-value (KV) cache of the prefixes. Therefore, in this paper, we explore the potential of prefix-sharing in the attention computation of the decode stage. However, the tree structure of the prefix-sharing mechanism presents significant challenges for attention computation in efficiently processing share"},"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":"2505.17694","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2025-05-23T10:03:28Z","cross_cats_sorted":[],"title_canon_sha256":"a2a0188f4d9caa42e7d0cbdb3c70141c8753a7271ac7607a12b26684fc4655bb","abstract_canon_sha256":"773f078215b15c0daa914f5b122e74e8bb8c2a16389f79c90053746ce9ab2da0"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-01T00:17:08.087005Z","signature_b64":"JnFZVnVsRAOWQN9ELQPNJrSMejjoRs87mKpUiKDicikDjOFbFbNUAvYzPV1rkVdphIYvY22WoAE2sbZHGakdAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"8623273386c752dd27674b7553b0cf381f431e43e8503d40a9540b3c2bffac09","last_reissued_at":"2026-07-01T00:17:08.086536Z","signature_status":"signed_v1","first_computed_at":"2026-07-01T00:17:08.086536Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"CoDec: Prefix-Shared Decoding Kernel for LLMs","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Chao Fang, Chengying Huan, Chen Tian, Guihai Chen, Kun Yang, Mo Zhou, Rong Gu, Rui Ning, Shaobo Ma, Sheng Zhong, Xi Lin, Xue Li, Zhibin Wang, Zhongfeng Wang, Zhonghui Zhang","submitted_at":"2025-05-23T10:03:28Z","abstract_excerpt":"Prefix-sharing among multiple prompts presents opportunities to combine the operations of the shared prefix, while attention computation in the decode stage, which becomes a critical bottleneck with increasing context lengths, is a memory-intensive process requiring heavy memory access on the key-value (KV) cache of the prefixes. Therefore, in this paper, we explore the potential of prefix-sharing in the attention computation of the decode stage. However, the tree structure of the prefix-sharing mechanism presents significant challenges for attention computation in efficiently processing share"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.17694","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/2505.17694/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":"2505.17694","created_at":"2026-07-01T00:17:08.086595+00:00"},{"alias_kind":"arxiv_version","alias_value":"2505.17694v2","created_at":"2026-07-01T00:17:08.086595+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.17694","created_at":"2026-07-01T00:17:08.086595+00:00"},{"alias_kind":"pith_short_12","alias_value":"QYRSOM4GY5JN","created_at":"2026-07-01T00:17:08.086595+00:00"},{"alias_kind":"pith_short_16","alias_value":"QYRSOM4GY5JN2J3H","created_at":"2026-07-01T00:17:08.086595+00:00"},{"alias_kind":"pith_short_8","alias_value":"QYRSOM4G","created_at":"2026-07-01T00:17:08.086595+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2606.07684","citing_title":"Semantic Cache Distillation: Efficient State Transfer via Reuse and Selective Patching","ref_index":54,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/QYRSOM4GY5JN2J3HJN2VHMGPHA","json":"https://pith.science/pith/QYRSOM4GY5JN2J3HJN2VHMGPHA.json","graph_json":"https://pith.science/api/pith-number/QYRSOM4GY5JN2J3HJN2VHMGPHA/graph.json","events_json":"https://pith.science/api/pith-number/QYRSOM4GY5JN2J3HJN2VHMGPHA/events.json","paper":"https://pith.science/paper/QYRSOM4G"},"agent_actions":{"view_html":"https://pith.science/pith/QYRSOM4GY5JN2J3HJN2VHMGPHA","download_json":"https://pith.science/pith/QYRSOM4GY5JN2J3HJN2VHMGPHA.json","view_paper":"https://pith.science/paper/QYRSOM4G","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2505.17694&json=true","fetch_graph":"https://pith.science/api/pith-number/QYRSOM4GY5JN2J3HJN2VHMGPHA/graph.json","fetch_events":"https://pith.science/api/pith-number/QYRSOM4GY5JN2J3HJN2VHMGPHA/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/QYRSOM4GY5JN2J3HJN2VHMGPHA/action/timestamp_anchor","attest_storage":"https://pith.science/pith/QYRSOM4GY5JN2J3HJN2VHMGPHA/action/storage_attestation","attest_author":"https://pith.science/pith/QYRSOM4GY5JN2J3HJN2VHMGPHA/action/author_attestation","sign_citation":"https://pith.science/pith/QYRSOM4GY5JN2J3HJN2VHMGPHA/action/citation_signature","submit_replication":"https://pith.science/pith/QYRSOM4GY5JN2J3HJN2VHMGPHA/action/replication_record"}},"created_at":"2026-07-01T00:17:08.086595+00:00","updated_at":"2026-07-01T00:17:08.086595+00:00"}