{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:AE6LMRNXCOQKWHDGUELLGETNXB","short_pith_number":"pith:AE6LMRNX","canonical_record":{"source":{"id":"2508.20900","kind":"arxiv","version":4},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.IR","submitted_at":"2025-08-28T15:29:51Z","cross_cats_sorted":[],"title_canon_sha256":"7d667165f61aa818b07c592cfb4e3d80281da910eea176e9b5143ed2b75e0261","abstract_canon_sha256":"c0facac20c3191a523b2f048357581392c86f5e99b9caf8abfc64735e63faefd"},"schema_version":"1.0"},"canonical_sha256":"013cb645b713a0ab1c66a116b3126db84c02bfe071884879c45f8b5485ed6f98","source":{"kind":"arxiv","id":"2508.20900","version":4},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2508.20900","created_at":"2026-05-17T23:38:46Z"},{"alias_kind":"arxiv_version","alias_value":"2508.20900v4","created_at":"2026-05-17T23:38:46Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2508.20900","created_at":"2026-05-17T23:38:46Z"},{"alias_kind":"pith_short_12","alias_value":"AE6LMRNXCOQK","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_16","alias_value":"AE6LMRNXCOQKWHDG","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_8","alias_value":"AE6LMRNX","created_at":"2026-05-18T12:33:37Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:AE6LMRNXCOQKWHDGUELLGETNXB","target":"record","payload":{"canonical_record":{"source":{"id":"2508.20900","kind":"arxiv","version":4},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.IR","submitted_at":"2025-08-28T15:29:51Z","cross_cats_sorted":[],"title_canon_sha256":"7d667165f61aa818b07c592cfb4e3d80281da910eea176e9b5143ed2b75e0261","abstract_canon_sha256":"c0facac20c3191a523b2f048357581392c86f5e99b9caf8abfc64735e63faefd"},"schema_version":"1.0"},"canonical_sha256":"013cb645b713a0ab1c66a116b3126db84c02bfe071884879c45f8b5485ed6f98","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-17T23:38:46.323537Z","signature_b64":"ce6JkQ2zprwLvO38XKAe6fa5hxHGmc1Fou2eeUpB/Pc+mA1/GTJ4st0nq7yUqzWucjPI3qW10GflnXO7GqH1CA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"013cb645b713a0ab1c66a116b3126db84c02bfe071884879c45f8b5485ed6f98","last_reissued_at":"2026-05-17T23:38:46.323034Z","signature_status":"signed_v1","first_computed_at":"2026-05-17T23:38:46.323034Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2508.20900","source_version":4,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-17T23:38:46Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"Zqf4rRHz6HebK0BAknuHPJSb60VWcpMOJXWIMwMIvtaEEc2RYsOO61M7Wf8AMLZipRsLSn89C2nxNj8+rywoDg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-30T07:47:35.433286Z"},"content_sha256":"100519603392c415dbe13ca7c6b293dd23f3ec76cb7332e398b9eba48c6c6ea0","schema_version":"1.0","event_id":"sha256:100519603392c415dbe13ca7c6b293dd23f3ec76cb7332e398b9eba48c6c6ea0"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:AE6LMRNXCOQKWHDGUELLGETNXB","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"OneRec-V2 Technical Report","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"Lazy decoder-only architecture cuts generative recommender computation by 94 percent and enables scaling to 8 billion parameters","cross_cats":[],"primary_cat":"cs.IR","authors_text":"Bin Zhang, Boxuan Wang, Chaoyi Ma, Chenglong Chu, Chengru Song, Chenhui Wang, Di Wang, Dongxue Meng, Dunju Zang, Fangyu Zhang, Fan Yang, Feng Jiang, Fuxing Zhang, Gang Wang, Guorui Zhou, Guowang Zhang, Han Li, Hengrui Hu, Honghui Bao, Hongtao Cheng, Hongyang Cao, Huanjie Wang, Jiaming Huang, Jiapeng Chen, Jiaqiang Liu, Jiaxin Deng, Jinghao Zhang, Jinghui Jia, Kun Gai, Kuo Cai, Lantao Hu, Lejian Ren, Liang Zeng, Liao Yu, Lu Ren, Pengfei Zheng, Qiang Luo, Qiang Wang, Qianqian Wang, Qidong Zhou, Qigen Hu, Rongzhou Zhang, Rui Huang, Ruiming Tang, Shengzhe Wang, Shihui He, Shiyao Wang, Shuang Yang, Shujie Yang, Siyang Mao, Sui Huang, Tao Wu, Tiantian He, Tingting Gao, Wei Yuan, Wuchao Li, Xiao Liang, Xiaoxiao Xu, Xinchen Luo, Xingmei Wang, Xugang Liu, Yang Zhou, Yan Wang, Yi Su, Yi Wang, Yiwu Liu, Yue Song, Yufei Zhang, Yunfan Wu, Yunfeng Zhao, Zexuan Cheng, Zhanyu Liu, Zhixin Ling, Ziming Li, Zixing Zhang","submitted_at":"2025-08-28T15:29:51Z","abstract_excerpt":"Recent breakthroughs in generative AI have transformed recommender systems through end-to-end generation. OneRec reformulates recommendation as an autoregressive generation task, achieving high Model FLOPs Utilization. While OneRec-V1 has shown significant empirical success in real-world deployment, two critical challenges hinder its scalability and performance: (1) inefficient computational allocation where 97.66% of resources are consumed by sequence encoding rather than generation, and (2) limitations in reinforcement learning relying solely on reward models.\n  To address these challenges, "},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"The Lazy Decoder-Only Architecture eliminates encoder bottlenecks, reducing total computation by 94% and training resources by 90%, enabling successful scaling to 8B parameters while Preference Alignment with Real-World User Interactions improves App Stay Time by 0.467%/0.741%.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That removing the encoder and relying on lazy decoding preserves recommendation quality and that duration-aware reward shaping plus adaptive clipping produce alignment that generalizes beyond the specific Kuaishou deployment without introducing new biases.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"OneRec-V2 scales generative recommendation to 8B parameters via decoder-only design and real-world preference alignment, improving user engagement metrics in production A/B tests.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Lazy decoder-only architecture cuts generative recommender computation by 94 percent and enables scaling to 8 billion parameters","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"44846a08951169ba24afdf08a4149392d39c832ac08a779d44b2343f8f83f092"},"source":{"id":"2508.20900","kind":"arxiv","version":4},"verdict":{"id":"8d4e5f5b-5b04-4f85-9910-0c659370c0b9","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-16T22:53:17.284049Z","strongest_claim":"The Lazy Decoder-Only Architecture eliminates encoder bottlenecks, reducing total computation by 94% and training resources by 90%, enabling successful scaling to 8B parameters while Preference Alignment with Real-World User Interactions improves App Stay Time by 0.467%/0.741%.","one_line_summary":"OneRec-V2 scales generative recommendation to 8B parameters via decoder-only design and real-world preference alignment, improving user engagement metrics in production A/B tests.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That removing the encoder and relying on lazy decoding preserves recommendation quality and that duration-aware reward shaping plus adaptive clipping produce alignment that generalizes beyond the specific Kuaishou deployment without introducing new biases.","pith_extraction_headline":"Lazy decoder-only architecture cuts generative recommender computation by 94 percent and enables scaling to 8 billion parameters"},"references":{"count":25,"sample":[{"doi":"","year":null,"title":"GPT-4 Technical Report","work_id":"b928e041-6991-4c08-8c81-0359e4097c7b","ref_index":1,"cited_arxiv_id":"2303.08774","is_internal_anchor":true},{"doi":"","year":null,"title":"GQA: Training Generalized Multi-Query Transformer Models from Multi-Head Checkpoints","work_id":"b73ad5b2-e553-4c71-b0c9-67e67ba7b158","ref_index":2,"cited_arxiv_id":"2305.13245","is_internal_anchor":true},{"doi":"","year":null,"title":"Pinrec: Outcome-conditioned, multi-token generative retrieval for industry-scale recommendation systems","work_id":"2dd1e9f2-dd74-4301-af2e-413e271f3aeb","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":1901,"title":"T. Brown, B. Mann, N. Ryder, M. Subbiah, J. D. Kaplan, P. Dhariwal, A. Neelakantan, P. Shyam, G. Sastry, A. Askell, et al. Language models are few-shot learners.Advances in neural information processi","work_id":"8e180b3a-a506-460e-b38b-5d27a2fb6ab2","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":null,"title":"MiniMax-M1: Scaling Test-Time Compute Efficiently with Lightning Attention","work_id":"c59fbe20-f41e-4140-a81c-40a12e7e8364","ref_index":5,"cited_arxiv_id":"2506.13585","is_internal_anchor":true}],"resolved_work":25,"snapshot_sha256":"6c44a69cc8bacc0993cfd2b1614aec83705051a73aca340ac6bfa0d1baa666ac","internal_anchors":14},"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"},"verdict_id":"8d4e5f5b-5b04-4f85-9910-0c659370c0b9"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-17T23:38:46Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"MPEvhkKT6Mf5fgnVdHq913vFqOCPQvOFgzMMT+OT+WMmhKFiwIlQ0z6XedRDVt67YVT65zeipB8StvMSaIW5Bw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-30T07:47:35.433839Z"},"content_sha256":"67efd4629753c3ec3011f44734350ff297d7454af11aae13185f7871c83a0705","schema_version":"1.0","event_id":"sha256:67efd4629753c3ec3011f44734350ff297d7454af11aae13185f7871c83a0705"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/AE6LMRNXCOQKWHDGUELLGETNXB/bundle.json","state_url":"https://pith.science/pith/AE6LMRNXCOQKWHDGUELLGETNXB/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/AE6LMRNXCOQKWHDGUELLGETNXB/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-05-30T07:47:35Z","links":{"resolver":"https://pith.science/pith/AE6LMRNXCOQKWHDGUELLGETNXB","bundle":"https://pith.science/pith/AE6LMRNXCOQKWHDGUELLGETNXB/bundle.json","state":"https://pith.science/pith/AE6LMRNXCOQKWHDGUELLGETNXB/state.json","well_known_bundle":"https://pith.science/.well-known/pith/AE6LMRNXCOQKWHDGUELLGETNXB/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:AE6LMRNXCOQKWHDGUELLGETNXB","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"c0facac20c3191a523b2f048357581392c86f5e99b9caf8abfc64735e63faefd","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.IR","submitted_at":"2025-08-28T15:29:51Z","title_canon_sha256":"7d667165f61aa818b07c592cfb4e3d80281da910eea176e9b5143ed2b75e0261"},"schema_version":"1.0","source":{"id":"2508.20900","kind":"arxiv","version":4}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2508.20900","created_at":"2026-05-17T23:38:46Z"},{"alias_kind":"arxiv_version","alias_value":"2508.20900v4","created_at":"2026-05-17T23:38:46Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2508.20900","created_at":"2026-05-17T23:38:46Z"},{"alias_kind":"pith_short_12","alias_value":"AE6LMRNXCOQK","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_16","alias_value":"AE6LMRNXCOQKWHDG","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_8","alias_value":"AE6LMRNX","created_at":"2026-05-18T12:33:37Z"}],"graph_snapshots":[{"event_id":"sha256:67efd4629753c3ec3011f44734350ff297d7454af11aae13185f7871c83a0705","target":"graph","created_at":"2026-05-17T23:38:46Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":4,"items":[{"attestation":"unclaimed","claim_id":"C1","kind":"strongest_claim","source":"verdict.strongest_claim","status":"machine_extracted","text":"The Lazy Decoder-Only Architecture eliminates encoder bottlenecks, reducing total computation by 94% and training resources by 90%, enabling successful scaling to 8B parameters while Preference Alignment with Real-World User Interactions improves App Stay Time by 0.467%/0.741%."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"That removing the encoder and relying on lazy decoding preserves recommendation quality and that duration-aware reward shaping plus adaptive clipping produce alignment that generalizes beyond the specific Kuaishou deployment without introducing new biases."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"OneRec-V2 scales generative recommendation to 8B parameters via decoder-only design and real-world preference alignment, improving user engagement metrics in production A/B tests."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"Lazy decoder-only architecture cuts generative recommender computation by 94 percent and enables scaling to 8 billion parameters"}],"snapshot_sha256":"44846a08951169ba24afdf08a4149392d39c832ac08a779d44b2343f8f83f092"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"paper":{"abstract_excerpt":"Recent breakthroughs in generative AI have transformed recommender systems through end-to-end generation. OneRec reformulates recommendation as an autoregressive generation task, achieving high Model FLOPs Utilization. While OneRec-V1 has shown significant empirical success in real-world deployment, two critical challenges hinder its scalability and performance: (1) inefficient computational allocation where 97.66% of resources are consumed by sequence encoding rather than generation, and (2) limitations in reinforcement learning relying solely on reward models.\n  To address these challenges, ","authors_text":"Bin Zhang, Boxuan Wang, Chaoyi Ma, Chenglong Chu, Chengru Song, Chenhui Wang, Di Wang, Dongxue Meng, Dunju Zang, Fangyu Zhang, Fan Yang, Feng Jiang, Fuxing Zhang, Gang Wang, Guorui Zhou, Guowang Zhang, Han Li, Hengrui Hu, Honghui Bao, Hongtao Cheng, Hongyang Cao, Huanjie Wang, Jiaming Huang, Jiapeng Chen, Jiaqiang Liu, Jiaxin Deng, Jinghao Zhang, Jinghui Jia, Kun Gai, Kuo Cai, Lantao Hu, Lejian Ren, Liang Zeng, Liao Yu, Lu Ren, Pengfei Zheng, Qiang Luo, Qiang Wang, Qianqian Wang, Qidong Zhou, Qigen Hu, Rongzhou Zhang, Rui Huang, Ruiming Tang, Shengzhe Wang, Shihui He, Shiyao Wang, Shuang Yang, Shujie Yang, Siyang Mao, Sui Huang, Tao Wu, Tiantian He, Tingting Gao, Wei Yuan, Wuchao Li, Xiao Liang, Xiaoxiao Xu, Xinchen Luo, Xingmei Wang, Xugang Liu, Yang Zhou, Yan Wang, Yi Su, Yi Wang, Yiwu Liu, Yue Song, Yufei Zhang, Yunfan Wu, Yunfeng Zhao, Zexuan Cheng, Zhanyu Liu, Zhixin Ling, Ziming Li, Zixing Zhang","cross_cats":[],"headline":"Lazy decoder-only architecture cuts generative recommender computation by 94 percent and enables scaling to 8 billion parameters","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.IR","submitted_at":"2025-08-28T15:29:51Z","title":"OneRec-V2 Technical Report"},"references":{"count":25,"internal_anchors":14,"resolved_work":25,"sample":[{"cited_arxiv_id":"2303.08774","doi":"","is_internal_anchor":true,"ref_index":1,"title":"GPT-4 Technical Report","work_id":"b928e041-6991-4c08-8c81-0359e4097c7b","year":null},{"cited_arxiv_id":"2305.13245","doi":"","is_internal_anchor":true,"ref_index":2,"title":"GQA: Training Generalized Multi-Query Transformer Models from Multi-Head Checkpoints","work_id":"b73ad5b2-e553-4c71-b0c9-67e67ba7b158","year":null},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":3,"title":"Pinrec: Outcome-conditioned, multi-token generative retrieval for industry-scale recommendation systems","work_id":"2dd1e9f2-dd74-4301-af2e-413e271f3aeb","year":null},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":4,"title":"T. Brown, B. Mann, N. Ryder, M. Subbiah, J. D. Kaplan, P. Dhariwal, A. Neelakantan, P. Shyam, G. Sastry, A. Askell, et al. Language models are few-shot learners.Advances in neural information processi","work_id":"8e180b3a-a506-460e-b38b-5d27a2fb6ab2","year":1901},{"cited_arxiv_id":"2506.13585","doi":"","is_internal_anchor":true,"ref_index":5,"title":"MiniMax-M1: Scaling Test-Time Compute Efficiently with Lightning Attention","work_id":"c59fbe20-f41e-4140-a81c-40a12e7e8364","year":null}],"snapshot_sha256":"6c44a69cc8bacc0993cfd2b1614aec83705051a73aca340ac6bfa0d1baa666ac"},"source":{"id":"2508.20900","kind":"arxiv","version":4},"verdict":{"created_at":"2026-05-16T22:53:17.284049Z","id":"8d4e5f5b-5b04-4f85-9910-0c659370c0b9","model_set":{"reader":"grok-4.3"},"one_line_summary":"OneRec-V2 scales generative recommendation to 8B parameters via decoder-only design and real-world preference alignment, improving user engagement metrics in production A/B tests.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"Lazy decoder-only architecture cuts generative recommender computation by 94 percent and enables scaling to 8 billion parameters","strongest_claim":"The Lazy Decoder-Only Architecture eliminates encoder bottlenecks, reducing total computation by 94% and training resources by 90%, enabling successful scaling to 8B parameters while Preference Alignment with Real-World User Interactions improves App Stay Time by 0.467%/0.741%.","weakest_assumption":"That removing the encoder and relying on lazy decoding preserves recommendation quality and that duration-aware reward shaping plus adaptive clipping produce alignment that generalizes beyond the specific Kuaishou deployment without introducing new biases."}},"verdict_id":"8d4e5f5b-5b04-4f85-9910-0c659370c0b9"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:100519603392c415dbe13ca7c6b293dd23f3ec76cb7332e398b9eba48c6c6ea0","target":"record","created_at":"2026-05-17T23:38:46Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"c0facac20c3191a523b2f048357581392c86f5e99b9caf8abfc64735e63faefd","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.IR","submitted_at":"2025-08-28T15:29:51Z","title_canon_sha256":"7d667165f61aa818b07c592cfb4e3d80281da910eea176e9b5143ed2b75e0261"},"schema_version":"1.0","source":{"id":"2508.20900","kind":"arxiv","version":4}},"canonical_sha256":"013cb645b713a0ab1c66a116b3126db84c02bfe071884879c45f8b5485ed6f98","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"013cb645b713a0ab1c66a116b3126db84c02bfe071884879c45f8b5485ed6f98","first_computed_at":"2026-05-17T23:38:46.323034Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-17T23:38:46.323034Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"ce6JkQ2zprwLvO38XKAe6fa5hxHGmc1Fou2eeUpB/Pc+mA1/GTJ4st0nq7yUqzWucjPI3qW10GflnXO7GqH1CA==","signature_status":"signed_v1","signed_at":"2026-05-17T23:38:46.323537Z","signed_message":"canonical_sha256_bytes"},"source_id":"2508.20900","source_kind":"arxiv","source_version":4}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:100519603392c415dbe13ca7c6b293dd23f3ec76cb7332e398b9eba48c6c6ea0","sha256:67efd4629753c3ec3011f44734350ff297d7454af11aae13185f7871c83a0705"],"state_sha256":"a080064a78b497a1b1e7413d48021bc1bd214f9615f8c8c96edae05a0a949418"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"CBgA89bPLRYUaBdGndk9hioirun9JFaLzQi6rRWxrdQH8YYxUOLbacpen00X8wTuRjd351PmKKGI68dHR/riDQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-05-30T07:47:35.436413Z","bundle_sha256":"67a394cac35ee42a57d07db6cfd7c794f5572aadb366c205684e6378baba94ae"}}