{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:37DRPZ725QH3NYKVGJ4NE63EUN","short_pith_number":"pith:37DRPZ72","canonical_record":{"source":{"id":"2409.13221","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.LG","submitted_at":"2024-09-20T05:15:38Z","cross_cats_sorted":["cs.CL","cs.DC"],"title_canon_sha256":"d7d48aacba83fb504e734055d88b5e4d4ab38facd86c6b35993168465d1ee4d8","abstract_canon_sha256":"244e3fa100c52507dab910133e9f98d337191efa72c9970ea12e33e8403acc9d"},"schema_version":"1.0"},"canonical_sha256":"dfc717e7faec0fb6e1553278d27b64a3400df3edfdb52e60b66a658c7b4227d2","source":{"kind":"arxiv","id":"2409.13221","version":3},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2409.13221","created_at":"2026-07-05T10:52:14Z"},{"alias_kind":"arxiv_version","alias_value":"2409.13221v3","created_at":"2026-07-05T10:52:14Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2409.13221","created_at":"2026-07-05T10:52:14Z"},{"alias_kind":"pith_short_12","alias_value":"37DRPZ725QH3","created_at":"2026-07-05T10:52:14Z"},{"alias_kind":"pith_short_16","alias_value":"37DRPZ725QH3NYKV","created_at":"2026-07-05T10:52:14Z"},{"alias_kind":"pith_short_8","alias_value":"37DRPZ72","created_at":"2026-07-05T10:52:14Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:37DRPZ725QH3NYKVGJ4NE63EUN","target":"record","payload":{"canonical_record":{"source":{"id":"2409.13221","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.LG","submitted_at":"2024-09-20T05:15:38Z","cross_cats_sorted":["cs.CL","cs.DC"],"title_canon_sha256":"d7d48aacba83fb504e734055d88b5e4d4ab38facd86c6b35993168465d1ee4d8","abstract_canon_sha256":"244e3fa100c52507dab910133e9f98d337191efa72c9970ea12e33e8403acc9d"},"schema_version":"1.0"},"canonical_sha256":"dfc717e7faec0fb6e1553278d27b64a3400df3edfdb52e60b66a658c7b4227d2","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:52:14.432801Z","signature_b64":"bQABmKsBz4NCDB9bHceUFlNwb8mKDulZ35SVIVL6hUBsgv2vUBsdJPHnZgquzVY+nAHUUW3EF7aD+pBKeLDdDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"dfc717e7faec0fb6e1553278d27b64a3400df3edfdb52e60b66a658c7b4227d2","last_reissued_at":"2026-07-05T10:52:14.432273Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:52:14.432273Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2409.13221","source_version":3,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T10:52:14Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"mG+L8mh+fJOP2K80G/yuVN0TkbUYHy9uZMC7855Owmt3oH7GuuBj4x12yRCfFRXf54/zhLllTy79wlRDC6UhAg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-18T17:10:44.026648Z"},"content_sha256":"bfd3548b8996dc656993647570847ea4589dc5615d5311ba79d95a257990ceda","schema_version":"1.0","event_id":"sha256:bfd3548b8996dc656993647570847ea4589dc5615d5311ba79d95a257990ceda"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:37DRPZ725QH3NYKVGJ4NE63EUN","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Optimizing RLHF Training for Large Language Models with Stage Fusion","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":["cs.CL","cs.DC"],"primary_cat":"cs.LG","authors_text":"Bingyang Wu, Changyi Wan, Hanpeng Hu, Lei Xia, Ranchen Ming, Shengyu Liu, Xin Jin, Yibo Zhu, Yinmin Zhong, Yukun Chen, Zili Zhang","submitted_at":"2024-09-20T05:15:38Z","abstract_excerpt":"We present RLHFuse, an efficient training system with stage fusion for Reinforcement Learning from Human Feedback (RLHF). Due to the intrinsic nature of RLHF training, i.e., the data skewness in the generation stage and the pipeline bubbles in the training stage, existing RLHF systems suffer from low GPU utilization. RLHFuse breaks the traditional view of RLHF workflow as a composition of individual tasks, splitting each task into finer-grained subtasks, and performing stage fusion to improve GPU utilization. RLHFuse contains two key ideas. First, for generation and inference tasks, RLHFuse sp"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2409.13221","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/2409.13221/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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T10:52:14Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"PmjA35hasa4xgsHZEmq4+jfF3r1Wuf/xUgGwCEAbV9cQWIWgDPooa9PK8leM35eDWfsXdbDn1Q/eyX0X03lcBA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-18T17:10:44.027003Z"},"content_sha256":"6c768a4e52e4b59b995ddd875195390255010cf6ae53cae81563df6bd33c562a","schema_version":"1.0","event_id":"sha256:6c768a4e52e4b59b995ddd875195390255010cf6ae53cae81563df6bd33c562a"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/37DRPZ725QH3NYKVGJ4NE63EUN/bundle.json","state_url":"https://pith.science/pith/37DRPZ725QH3NYKVGJ4NE63EUN/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/37DRPZ725QH3NYKVGJ4NE63EUN/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":"2026-08-18T17:10:44Z","links":{"resolver":"https://pith.science/pith/37DRPZ725QH3NYKVGJ4NE63EUN","bundle":"https://pith.science/pith/37DRPZ725QH3NYKVGJ4NE63EUN/bundle.json","state":"https://pith.science/pith/37DRPZ725QH3NYKVGJ4NE63EUN/state.json","well_known_bundle":"https://pith.science/.well-known/pith/37DRPZ725QH3NYKVGJ4NE63EUN/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:37DRPZ725QH3NYKVGJ4NE63EUN","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":"244e3fa100c52507dab910133e9f98d337191efa72c9970ea12e33e8403acc9d","cross_cats_sorted":["cs.CL","cs.DC"],"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.LG","submitted_at":"2024-09-20T05:15:38Z","title_canon_sha256":"d7d48aacba83fb504e734055d88b5e4d4ab38facd86c6b35993168465d1ee4d8"},"schema_version":"1.0","source":{"id":"2409.13221","kind":"arxiv","version":3}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2409.13221","created_at":"2026-07-05T10:52:14Z"},{"alias_kind":"arxiv_version","alias_value":"2409.13221v3","created_at":"2026-07-05T10:52:14Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2409.13221","created_at":"2026-07-05T10:52:14Z"},{"alias_kind":"pith_short_12","alias_value":"37DRPZ725QH3","created_at":"2026-07-05T10:52:14Z"},{"alias_kind":"pith_short_16","alias_value":"37DRPZ725QH3NYKV","created_at":"2026-07-05T10:52:14Z"},{"alias_kind":"pith_short_8","alias_value":"37DRPZ72","created_at":"2026-07-05T10:52:14Z"}],"graph_snapshots":[{"event_id":"sha256:6c768a4e52e4b59b995ddd875195390255010cf6ae53cae81563df6bd33c562a","target":"graph","created_at":"2026-07-05T10:52:14Z","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":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2409.13221/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"We present RLHFuse, an efficient training system with stage fusion for Reinforcement Learning from Human Feedback (RLHF). Due to the intrinsic nature of RLHF training, i.e., the data skewness in the generation stage and the pipeline bubbles in the training stage, existing RLHF systems suffer from low GPU utilization. RLHFuse breaks the traditional view of RLHF workflow as a composition of individual tasks, splitting each task into finer-grained subtasks, and performing stage fusion to improve GPU utilization. RLHFuse contains two key ideas. First, for generation and inference tasks, RLHFuse sp","authors_text":"Bingyang Wu, Changyi Wan, Hanpeng Hu, Lei Xia, Ranchen Ming, Shengyu Liu, Xin Jin, Yibo Zhu, Yinmin Zhong, Yukun Chen, Zili Zhang","cross_cats":["cs.CL","cs.DC"],"headline":"","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.LG","submitted_at":"2024-09-20T05:15:38Z","title":"Optimizing RLHF Training for Large Language Models with Stage Fusion"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2409.13221","kind":"arxiv","version":3},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:bfd3548b8996dc656993647570847ea4589dc5615d5311ba79d95a257990ceda","target":"record","created_at":"2026-07-05T10:52:14Z","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":"244e3fa100c52507dab910133e9f98d337191efa72c9970ea12e33e8403acc9d","cross_cats_sorted":["cs.CL","cs.DC"],"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.LG","submitted_at":"2024-09-20T05:15:38Z","title_canon_sha256":"d7d48aacba83fb504e734055d88b5e4d4ab38facd86c6b35993168465d1ee4d8"},"schema_version":"1.0","source":{"id":"2409.13221","kind":"arxiv","version":3}},"canonical_sha256":"dfc717e7faec0fb6e1553278d27b64a3400df3edfdb52e60b66a658c7b4227d2","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"dfc717e7faec0fb6e1553278d27b64a3400df3edfdb52e60b66a658c7b4227d2","first_computed_at":"2026-07-05T10:52:14.432273Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T10:52:14.432273Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"bQABmKsBz4NCDB9bHceUFlNwb8mKDulZ35SVIVL6hUBsgv2vUBsdJPHnZgquzVY+nAHUUW3EF7aD+pBKeLDdDQ==","signature_status":"signed_v1","signed_at":"2026-07-05T10:52:14.432801Z","signed_message":"canonical_sha256_bytes"},"source_id":"2409.13221","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:bfd3548b8996dc656993647570847ea4589dc5615d5311ba79d95a257990ceda","sha256:6c768a4e52e4b59b995ddd875195390255010cf6ae53cae81563df6bd33c562a"],"state_sha256":"8e540c566aeac9770e38ed6f96fced914f6d60ec668978647868853eba23f564"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"h5v8iK2rP4W1H7yQiqs+OxqViq1V15j31VZDIupnVzARmi7VewhUwYxRORlKOqC19+/HuSbMIEpUDqLE/19/Cg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-18T17:10:44.030285Z","bundle_sha256":"ac264aa804370151452ab9a50c771966ed124cf320363f9a67b78961e299354f"}}