{"as_of":"2026-08-09T14:06:00Z","caps":{"database_statements":6,"inbound":100,"outbound":100},"context_digest":"sha256:b77b4d1e49bf7c2118d2e6c35edea53d6e2e6eb2e5ed718b14f21f86a47b90ac","coverage":[{"denominator":41,"lane":"reference_resolution","note":"Typed states for the displayed outbound observations.","records_observed":41,"source":"paper_references, paper_reference_links","source_observed_at":"2026-08-02T11:14:02.893027Z","state":"measured"},{"denominator":41,"lane":"standing_notices","note":"One-hop event checks from named stored sources.","records_observed":41,"source":"scholarly_work_events, retraction_status_cache","source_observed_at":"2026-08-09T06:31:02.800959+00:00","state":"measured"},{"denominator":0,"lane":"inbound_itemization","note":"Pith citing papers itemized under the disclosed page cap.","records_observed":0,"source":"paper_references, paper_reference_links","source_observed_at":null,"state":"measured"},{"denominator":1,"lane":"external_citation_measurements","note":"A source-named dated measurement, never combined with another source.","records_observed":0,"source":"cited_works","source_observed_at":null,"state":"measured"}],"external_citation_measurements":[],"inbound":[],"links":{"evidence":"/evidence","html":"/paper/2607.22614/citation-record","integrity":"/paper/2607.22614/integrity","json":"/paper/2607.22614/citation-record.json","paper":"/paper/2607.22614"},"outbound":[{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-02T11:13:59.202663Z","title":"2025.Training-Free Group Relative Policy Optimization","venue":null,"work_id":null,"year":2025},"citing_paper":{"arxiv_id":"2607.22614","last_updated":"2026-06-15T09:41:09Z","snapshot_observed_at":"2026-08-07T04:23:01.926858Z","submitted_at":"2026-06-15T09:41:09Z","title":"DynaResize: Runtime GPU Reallocation for Disaggregated LLM Post-Training","version":1},"reference_index":1,"source":"pdf_text","source_observed_at":"2026-08-02T11:13:59.202663Z"},"links":{"citing_paper":"/paper/2607.22614"},"observation_digest":"sha256:14c1d16fb9e90ac0b4d0e92b1c51d9f117e2e46f0f78510eb7b2f7f6f6b12154","observation_id":"431f9042-c323-4147-8be6-22b5af8328c0","resolution":{"observed_at":"2026-08-02T11:13:59.202663Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-02T11:13:59.239053Z","title":null,"venue":null,"work_id":null,"year":2026},"citing_paper":{"arxiv_id":"2607.22614","last_updated":"2026-06-15T09:41:09Z","snapshot_observed_at":"2026-08-07T04:23:01.926858Z","submitted_at":"2026-06-15T09:41:09Z","title":"DynaResize: Runtime GPU Reallocation for Disaggregated LLM Post-Training","version":1},"reference_index":2,"source":"pdf_text","source_observed_at":"2026-08-02T11:13:59.239053Z"},"links":{"citing_paper":"/paper/2607.22614"},"observation_digest":"sha256:40732174d84416fa675f2db38f9c90861d9b2fe8a5130ad11afccb50a7332fea","observation_id":"a579108a-d7d1-4102-84e2-6944a3776cde","resolution":{"observed_at":"2026-08-02T11:13:59.239053Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-02T11:13:59.364418Z","title":null,"venue":null,"work_id":null,"year":2026},"citing_paper":{"arxiv_id":"2607.22614","last_updated":"2026-06-15T09:41:09Z","snapshot_observed_at":"2026-08-07T04:23:01.926858Z","submitted_at":"2026-06-15T09:41:09Z","title":"DynaResize: Runtime GPU Reallocation for Disaggregated LLM Post-Training","version":1},"reference_index":3,"source":"pdf_text","source_observed_at":"2026-08-02T11:13:59.364418Z"},"links":{"citing_paper":"/paper/2607.22614"},"observation_digest":"sha256:0b265338d100ff13a60b09d030e3ce1e3afa1021ae766ead2cb05f8ba10b0453","observation_id":"3a92ece5-3fd6-4962-9ea1-44fe1a5ab89b","resolution":{"observed_at":"2026-08-02T11:13:59.364418Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2407.21783","last_updated":"2024-11-23T23:27:33Z","snapshot_observed_at":"2026-07-06T18:55:11.576666Z","submitted_at":"2024-07-31T17:54:27Z","title":"The Llama 3 Herd of Models","version":3},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2407.21783","snapshot_observed_at":"2026-08-02T11:13:59.448930Z","title":"2024.The Llama 3 Herd of Models","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2607.22614","last_updated":"2026-06-15T09:41:09Z","snapshot_observed_at":"2026-08-07T04:23:01.926858Z","submitted_at":"2026-06-15T09:41:09Z","title":"DynaResize: Runtime GPU Reallocation for Disaggregated LLM Post-Training","version":1},"reference_index":4,"source":"pdf_text","source_observed_at":"2026-08-02T11:13:59.448930Z"},"links":{"cited_paper":"/paper/2407.21783","citing_paper":"/paper/2607.22614"},"observation_digest":"sha256:405c4535b80cfab7c63d90f3e4aec20e80715f9c883c1bb163b129e2f1a8a78c","observation_id":"0615b364-8985-429f-b35d-ff598d61b3a2","resolution":{"observed_at":"2026-08-02T11:13:59.448930Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2506.15050","last_updated":"2025-06-18T01:21:38Z","snapshot_observed_at":"2026-08-07T00:00:11.300845Z","submitted_at":"2025-06-18T01:21:38Z","title":"Truncated Proximal Policy Optimization","version":1},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2506.15050","snapshot_observed_at":"2026-08-02T11:13:59.500195Z","title":"2025.Truncated Proximal Policy Optimization","venue":null,"work_id":null,"year":2025},"citing_paper":{"arxiv_id":"2607.22614","last_updated":"2026-06-15T09:41:09Z","snapshot_observed_at":"2026-08-07T04:23:01.926858Z","submitted_at":"2026-06-15T09:41:09Z","title":"DynaResize: Runtime GPU Reallocation for Disaggregated LLM Post-Training","version":1},"reference_index":5,"source":"pdf_text","source_observed_at":"2026-08-02T11:13:59.500195Z"},"links":{"cited_paper":"/paper/2506.15050","citing_paper":"/paper/2607.22614"},"observation_digest":"sha256:0e899e64075058fa61e7a2da7af93bd43b4b18ef2ca580f19cb36bf8cb954234","observation_id":"64ffcc2a-3cd9-4eea-bcba-de093c193ae0","resolution":{"observed_at":"2026-08-02T11:13:59.500195Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2505.24298","last_updated":"2026-03-02T03:07:29Z","snapshot_observed_at":"2026-08-02T16:10:09.199211Z","submitted_at":"2025-05-30T07:18:25Z","title":"AReaL: A Large-Scale Asynchronous Reinforcement Learning System for Language Reasoning","version":5},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2505.24298","snapshot_observed_at":"2026-08-02T11:13:59.560465Z","title":null,"venue":null,"work_id":null,"year":2025},"citing_paper":{"arxiv_id":"2607.22614","last_updated":"2026-06-15T09:41:09Z","snapshot_observed_at":"2026-08-07T04:23:01.926858Z","submitted_at":"2026-06-15T09:41:09Z","title":"DynaResize: Runtime GPU Reallocation for Disaggregated LLM Post-Training","version":1},"reference_index":6,"source":"pdf_text","source_observed_at":"2026-08-02T11:13:59.560465Z"},"links":{"cited_paper":"/paper/2505.24298","citing_paper":"/paper/2607.22614"},"observation_digest":"sha256:965d07e3dd9eb002d78bc110b670784c80e60ad4ca2dede496f625b3e563f9c2","observation_id":"4a1c46ae-723a-478a-990e-bf9e9890cca4","resolution":{"observed_at":"2026-08-02T11:13:59.560465Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-02T11:13:59.614113Z","title":null,"venue":null,"work_id":null,"year":2026},"citing_paper":{"arxiv_id":"2607.22614","last_updated":"2026-06-15T09:41:09Z","snapshot_observed_at":"2026-08-07T04:23:01.926858Z","submitted_at":"2026-06-15T09:41:09Z","title":"DynaResize: Runtime GPU Reallocation for Disaggregated LLM Post-Training","version":1},"reference_index":7,"source":"pdf_text","source_observed_at":"2026-08-02T11:13:59.614113Z"},"links":{"citing_paper":"/paper/2607.22614"},"observation_digest":"sha256:3a11cefaf62d2bfa3bb3f45af37537e6512571e3679bd844fa8895a03de41c1e","observation_id":"934a0579-dea5-41cb-89e3-6eddc2b36e16","resolution":{"observed_at":"2026-08-02T11:13:59.614113Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2605.06534","last_updated":"2026-05-20T11:37:51Z","snapshot_observed_at":"2026-08-03T06:43:16.351179Z","submitted_at":"2026-05-07T16:33:40Z","title":"ROSE: Rollout On Serving GPUs via Cooperative Elasticity for Agentic RL","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2605.06534","snapshot_observed_at":"2026-08-02T11:13:59.715326Z","title":"2026.ROSE: Rollout On Serving GPUs via Cooperative Elasticity for Agentic RL","venue":null,"work_id":null,"year":2026},"citing_paper":{"arxiv_id":"2607.22614","last_updated":"2026-06-15T09:41:09Z","snapshot_observed_at":"2026-08-07T04:23:01.926858Z","submitted_at":"2026-06-15T09:41:09Z","title":"DynaResize: Runtime GPU Reallocation for Disaggregated LLM Post-Training","version":1},"reference_index":8,"source":"pdf_text","source_observed_at":"2026-08-02T11:13:59.715326Z"},"links":{"cited_paper":"/paper/2605.06534","citing_paper":"/paper/2607.22614"},"observation_digest":"sha256:7c2355b0eb951fb42385a164f28d1c35d850fda8b9585883b2f4701c17e1b6ad","observation_id":"a36a1558-768b-40de-9ae5-7aceb0913ee1","resolution":{"observed_at":"2026-08-02T11:13:59.715326Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2412.04504","last_updated":"2024-12-03T03:16:12Z","snapshot_observed_at":"2026-08-09T13:13:40.709458Z","submitted_at":"2024-12-03T03:16:12Z","title":"Multi-Bin Batching for Increasing LLM Inference Throughput","version":1},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2412.04504","snapshot_observed_at":"2026-08-02T11:13:59.783585Z","title":"2024.Multi-Bin Batching for Increasing LLM Inference Throughput","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2607.22614","last_updated":"2026-06-15T09:41:09Z","snapshot_observed_at":"2026-08-07T04:23:01.926858Z","submitted_at":"2026-06-15T09:41:09Z","title":"DynaResize: Runtime GPU Reallocation for Disaggregated LLM Post-Training","version":1},"reference_index":9,"source":"pdf_text","source_observed_at":"2026-08-02T11:13:59.783585Z"},"links":{"cited_paper":"/paper/2412.04504","citing_paper":"/paper/2607.22614"},"observation_digest":"sha256:721227dfbfa503f2a9ec32873f64fe5e722edafdbbf57580213106600fc7e250","observation_id":"baa5ca5c-598f-4937-ba7a-7c0a3fb3e31a","resolution":{"observed_at":"2026-08-02T11:13:59.783585Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2507.01663","last_updated":"2025-07-02T12:45:34Z","snapshot_observed_at":"2026-08-09T04:10:37.358894Z","submitted_at":"2025-07-02T12:45:34Z","title":"AsyncFlow: An Asynchronous Streaming RL Framework for Efficient LLM Post-Training","version":1},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2507.01663","snapshot_observed_at":"2026-08-02T11:13:59.835182Z","title":"2025.AsyncFlow: An Asynchronous Streaming RL Framework for Efficient LLM Post-Training","venue":null,"work_id":null,"year":2025},"citing_paper":{"arxiv_id":"2607.22614","last_updated":"2026-06-15T09:41:09Z","snapshot_observed_at":"2026-08-07T04:23:01.926858Z","submitted_at":"2026-06-15T09:41:09Z","title":"DynaResize: Runtime GPU Reallocation for Disaggregated LLM Post-Training","version":1},"reference_index":10,"source":"pdf_text","source_observed_at":"2026-08-02T11:13:59.835182Z"},"links":{"cited_paper":"/paper/2507.01663","citing_paper":"/paper/2607.22614"},"observation_digest":"sha256:2c75ac05675ded05f0392b87f63b881383df10d9d44d3a70d2d1ce7657b25663","observation_id":"b2a7c890-de8e-43d3-bd39-5f0139bbfee1","resolution":{"observed_at":"2026-08-02T11:13:59.835182Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-02T11:13:59.883134Z","title":null,"venue":null,"work_id":null,"year":2023},"citing_paper":{"arxiv_id":"2607.22614","last_updated":"2026-06-15T09:41:09Z","snapshot_observed_at":"2026-08-07T04:23:01.926858Z","submitted_at":"2026-06-15T09:41:09Z","title":"DynaResize: Runtime GPU Reallocation for Disaggregated LLM Post-Training","version":1},"reference_index":11,"source":"pdf_text","source_observed_at":"2026-08-02T11:13:59.883134Z"},"links":{"citing_paper":"/paper/2607.22614"},"observation_digest":"sha256:1ac37adb700b372138e5ee23c397038efd18fb22d15f8444458ab7c8b3921dc7","observation_id":"4b1b6617-ecc0-486d-ae53-b3a0e3134f96","resolution":{"observed_at":"2026-08-02T11:13:59.883134Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2309.06180","last_updated":"2023-09-12T12:50:04Z","snapshot_observed_at":"2026-08-02T09:51:08.145755Z","submitted_at":"2023-09-12T12:50:04Z","title":"Efficient Memory Management for Large Language Model Serving with PagedAttention","version":1},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2309.06180","snapshot_observed_at":"2026-08-02T11:13:59.967945Z","title":"Gonzalez, Hao Zhang, and Ion Stoica","venue":null,"work_id":null,"year":2023},"citing_paper":{"arxiv_id":"2607.22614","last_updated":"2026-06-15T09:41:09Z","snapshot_observed_at":"2026-08-07T04:23:01.926858Z","submitted_at":"2026-06-15T09:41:09Z","title":"DynaResize: Runtime GPU Reallocation for Disaggregated LLM Post-Training","version":1},"reference_index":12,"source":"pdf_text","source_observed_at":"2026-08-02T11:13:59.967945Z"},"links":{"cited_paper":"/paper/2309.06180","citing_paper":"/paper/2607.22614"},"observation_digest":"sha256:65659d6268970340e1025ea946b2d9027b3cbefcc94eaae5dc088bf7a9305345","observation_id":"69b1ceb7-c6eb-4321-98db-b95fef3aa2b0","resolution":{"observed_at":"2026-08-02T11:13:59.967945Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2406.14088","last_updated":"2025-04-24T13:24:42Z","snapshot_observed_at":"2026-08-09T10:40:45.328760Z","submitted_at":"2024-06-20T08:04:07Z","title":"ReaL: Efficient RLHF Training of Large Language Models with Parameter Reallocation","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2406.14088","snapshot_observed_at":"2026-08-02T11:14:00.088095Z","title":"2024.ReaL: Efficient RLHF Training of Large Language Models with Parameter Reallocation","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2607.22614","last_updated":"2026-06-15T09:41:09Z","snapshot_observed_at":"2026-08-07T04:23:01.926858Z","submitted_at":"2026-06-15T09:41:09Z","title":"DynaResize: Runtime GPU Reallocation for Disaggregated LLM Post-Training","version":1},"reference_index":13,"source":"pdf_text","source_observed_at":"2026-08-02T11:14:00.088095Z"},"links":{"cited_paper":"/paper/2406.14088","citing_paper":"/paper/2607.22614"},"observation_digest":"sha256:7e084863bad7d00c75e3e8b665d4ccf7b54eba1174f626a6eeea52b71cc518eb","observation_id":"c920d508-64bc-4f6d-b9d8-261ba752fd9e","resolution":{"observed_at":"2026-08-02T11:14:00.088095Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-02T11:14:00.236726Z","title":"Jordan, and Ion Stoica","venue":null,"work_id":null,"year":2018},"citing_paper":{"arxiv_id":"2607.22614","last_updated":"2026-06-15T09:41:09Z","snapshot_observed_at":"2026-08-07T04:23:01.926858Z","submitted_at":"2026-06-15T09:41:09Z","title":"DynaResize: Runtime GPU Reallocation for Disaggregated LLM Post-Training","version":1},"reference_index":14,"source":"pdf_text","source_observed_at":"2026-08-02T11:14:00.236726Z"},"links":{"citing_paper":"/paper/2607.22614"},"observation_digest":"sha256:77ec84a3b698529ab38bd6b2428fa57a8d4218cb7b9168ee55b98d638c10a021","observation_id":"9363794a-c91b-4bd0-a8cd-c42d2c498d46","resolution":{"observed_at":"2026-08-02T11:14:00.236726Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-02T11:14:00.355894Z","title":null,"venue":null,"work_id":null,"year":2026},"citing_paper":{"arxiv_id":"2607.22614","last_updated":"2026-06-15T09:41:09Z","snapshot_observed_at":"2026-08-07T04:23:01.926858Z","submitted_at":"2026-06-15T09:41:09Z","title":"DynaResize: Runtime GPU Reallocation for Disaggregated LLM Post-Training","version":1},"reference_index":15,"source":"pdf_text","source_observed_at":"2026-08-02T11:14:00.355894Z"},"links":{"citing_paper":"/paper/2607.22614"},"observation_digest":"sha256:adc4b50d4df5170153da008cec679607f9745b1a43da6a22c3be879df48c76a9","observation_id":"800d9ab0-6c63-49a5-b5c0-cf0998cda45f","resolution":{"observed_at":"2026-08-02T11:14:00.355894Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-02T11:14:00.489562Z","title":"Or et al","venue":null,"work_id":null,"year":2020},"citing_paper":{"arxiv_id":"2607.22614","last_updated":"2026-06-15T09:41:09Z","snapshot_observed_at":"2026-08-07T04:23:01.926858Z","submitted_at":"2026-06-15T09:41:09Z","title":"DynaResize: Runtime GPU Reallocation for Disaggregated LLM Post-Training","version":1},"reference_index":16,"source":"pdf_text","source_observed_at":"2026-08-02T11:14:00.489562Z"},"links":{"citing_paper":"/paper/2607.22614"},"observation_digest":"sha256:b2ed63f30524c84ccdbd6f09bfe3cc62e18900df5a047aa0622664369ccd8aec","observation_id":"52528263-6885-4fa1-aa06-0faa8c3ed7ae","resolution":{"observed_at":"2026-08-02T11:14:00.489562Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-02T11:14:00.610430Z","title":null,"venue":null,"work_id":null,"year":2022},"citing_paper":{"arxiv_id":"2607.22614","last_updated":"2026-06-15T09:41:09Z","snapshot_observed_at":"2026-08-07T04:23:01.926858Z","submitted_at":"2026-06-15T09:41:09Z","title":"DynaResize: Runtime GPU Reallocation for Disaggregated LLM Post-Training","version":1},"reference_index":17,"source":"pdf_text","source_observed_at":"2026-08-02T11:14:00.610430Z"},"links":{"citing_paper":"/paper/2607.22614"},"observation_digest":"sha256:75b8ec8011e01da604b81209deb2815c965ff47c0689cab8aa670f8d76a5a3b4","observation_id":"467d101b-13aa-4c6b-a590-fc093c2139b2","resolution":{"observed_at":"2026-08-02T11:14:00.610430Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-02T11:14:00.792222Z","title":null,"venue":null,"work_id":null,"year":2019},"citing_paper":{"arxiv_id":"2607.22614","last_updated":"2026-06-15T09:41:09Z","snapshot_observed_at":"2026-08-07T04:23:01.926858Z","submitted_at":"2026-06-15T09:41:09Z","title":"DynaResize: Runtime GPU Reallocation for Disaggregated LLM Post-Training","version":1},"reference_index":18,"source":"pdf_text","source_observed_at":"2026-08-02T11:14:00.792222Z"},"links":{"citing_paper":"/paper/2607.22614"},"observation_digest":"sha256:ff8a4ac1bda0383f19244bc4d2a9fa40c5b63911f3c5c31654e88badd96c7b1e","observation_id":"636b03fe-871f-4f16-bdce-22cf65cc0d47","resolution":{"observed_at":"2026-08-02T11:14:00.792222Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-02T11:14:00.885566Z","title":null,"venue":null,"work_id":null,"year":2018},"citing_paper":{"arxiv_id":"2607.22614","last_updated":"2026-06-15T09:41:09Z","snapshot_observed_at":"2026-08-07T04:23:01.926858Z","submitted_at":"2026-06-15T09:41:09Z","title":"DynaResize: Runtime GPU Reallocation for Disaggregated LLM Post-Training","version":1},"reference_index":19,"source":"pdf_text","source_observed_at":"2026-08-02T11:14:00.885566Z"},"links":{"citing_paper":"/paper/2607.22614"},"observation_digest":"sha256:b5f02210c77c3f33c3d5302bb145236d037e46c00ef6bb5122094d49d5fb65cc","observation_id":"7bb1c0f3-0cd4-4b14-8fe8-d3314c7601b5","resolution":{"observed_at":"2026-08-02T11:14:00.885566Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-02T11:14:00.959831Z","title":null,"venue":null,"work_id":null,"year":2023},"citing_paper":{"arxiv_id":"2607.22614","last_updated":"2026-06-15T09:41:09Z","snapshot_observed_at":"2026-08-07T04:23:01.926858Z","submitted_at":"2026-06-15T09:41:09Z","title":"DynaResize: Runtime GPU Reallocation for Disaggregated LLM Post-Training","version":1},"reference_index":20,"source":"pdf_text","source_observed_at":"2026-08-02T11:14:00.959831Z"},"links":{"citing_paper":"/paper/2607.22614"},"observation_digest":"sha256:fbea093d8406bfd5833d591078350d4e8f96e9551fa273969b810bdbcb47e974","observation_id":"203322d9-e1de-4299-938e-2b10dd30ebc5","resolution":{"observed_at":"2026-08-02T11:14:00.959831Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-02T11:14:01.005496Z","title":"Rajbhandari et al","venue":null,"work_id":null,"year":2020},"citing_paper":{"arxiv_id":"2607.22614","last_updated":"2026-06-15T09:41:09Z","snapshot_observed_at":"2026-08-07T04:23:01.926858Z","submitted_at":"2026-06-15T09:41:09Z","title":"DynaResize: Runtime GPU Reallocation for Disaggregated LLM Post-Training","version":1},"reference_index":21,"source":"pdf_text","source_observed_at":"2026-08-02T11:14:01.005496Z"},"links":{"citing_paper":"/paper/2607.22614"},"observation_digest":"sha256:256d1e16a222e0106963446410b6804475b9809524995b7ad7b5face6836b9f9","observation_id":"3ab533a5-50d8-445b-92df-50f0f278bf79","resolution":{"observed_at":"2026-08-02T11:14:01.005496Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-02T11:14:01.104928Z","title":null,"venue":null,"work_id":null,"year":2021},"citing_paper":{"arxiv_id":"2607.22614","last_updated":"2026-06-15T09:41:09Z","snapshot_observed_at":"2026-08-07T04:23:01.926858Z","submitted_at":"2026-06-15T09:41:09Z","title":"DynaResize: Runtime GPU Reallocation for Disaggregated LLM Post-Training","version":1},"reference_index":22,"source":"pdf_text","source_observed_at":"2026-08-02T11:14:01.104928Z"},"links":{"citing_paper":"/paper/2607.22614"},"observation_digest":"sha256:69efd39f7e03ba60c8f60e270f6e635bf744205f81d240aed29048937684605f","observation_id":"e1ae4feb-551d-42a5-bce0-cf652054424e","resolution":{"observed_at":"2026-08-02T11:14:01.104928Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-02T11:14:01.188380Z","title":null,"venue":null,"work_id":null,"year":2021},"citing_paper":{"arxiv_id":"2607.22614","last_updated":"2026-06-15T09:41:09Z","snapshot_observed_at":"2026-08-07T04:23:01.926858Z","submitted_at":"2026-06-15T09:41:09Z","title":"DynaResize: Runtime GPU Reallocation for Disaggregated LLM Post-Training","version":1},"reference_index":23,"source":"pdf_text","source_observed_at":"2026-08-02T11:14:01.188380Z"},"links":{"citing_paper":"/paper/2607.22614"},"observation_digest":"sha256:5bbd4a16d672f9145ac38765a3230ee8df746d8186ac50efc0dfd743a843c82a","observation_id":"ef1bd6c9-7f1a-4fae-9c84-da8c1dc64402","resolution":{"observed_at":"2026-08-02T11:14:01.188380Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"1707.06347","last_updated":"2017-08-28T09:20:06Z","snapshot_observed_at":"2026-07-06T02:11:23.670680Z","submitted_at":"2017-07-20T02:32:33Z","title":"Proximal Policy Optimization Algorithms","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"1707.06347","snapshot_observed_at":"2026-08-02T11:14:01.264828Z","title":"2017.Proximal policy optimization algorithms","venue":null,"work_id":null,"year":2017},"citing_paper":{"arxiv_id":"2607.22614","last_updated":"2026-06-15T09:41:09Z","snapshot_observed_at":"2026-08-07T04:23:01.926858Z","submitted_at":"2026-06-15T09:41:09Z","title":"DynaResize: Runtime GPU Reallocation for Disaggregated LLM Post-Training","version":1},"reference_index":24,"source":"pdf_text","source_observed_at":"2026-08-02T11:14:01.264828Z"},"links":{"cited_paper":"/paper/1707.06347","citing_paper":"/paper/2607.22614"},"observation_digest":"sha256:356afe545b3f4a13c07eabab0cb8477eee0c5f35b2424db5d1c7c90bb391f706","observation_id":"3a025a21-d411-4d74-9349-f99f7f0bdeb7","resolution":{"observed_at":"2026-08-02T11:14:01.264828Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-02T11:14:01.374900Z","title":null,"venue":null,"work_id":null,"year":2026},"citing_paper":{"arxiv_id":"2607.22614","last_updated":"2026-06-15T09:41:09Z","snapshot_observed_at":"2026-08-07T04:23:01.926858Z","submitted_at":"2026-06-15T09:41:09Z","title":"DynaResize: Runtime GPU Reallocation for Disaggregated LLM Post-Training","version":1},"reference_index":25,"source":"pdf_text","source_observed_at":"2026-08-02T11:14:01.374900Z"},"links":{"citing_paper":"/paper/2607.22614"},"observation_digest":"sha256:98dfae74c3b5dbff5a166d8770f3cf5f783065492a99295e52ff1a30cffd7df8","observation_id":"93699b67-0772-472d-ae34-ee3f2b1d16ff","resolution":{"observed_at":"2026-08-02T11:14:01.374900Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-02T11:14:01.494834Z","title":null,"venue":null,"work_id":null,"year":2025},"citing_paper":{"arxiv_id":"2607.22614","last_updated":"2026-06-15T09:41:09Z","snapshot_observed_at":"2026-08-07T04:23:01.926858Z","submitted_at":"2026-06-15T09:41:09Z","title":"DynaResize: Runtime GPU Reallocation for Disaggregated LLM Post-Training","version":1},"reference_index":26,"source":"pdf_text","source_observed_at":"2026-08-02T11:14:01.494834Z"},"links":{"citing_paper":"/paper/2607.22614"},"observation_digest":"sha256:2442684180f5e7543a2bafd24038443161663761bd0bde2964c0a74a72e458b7","observation_id":"67a18c02-0280-49ca-b842-6973ac330add","resolution":{"observed_at":"2026-08-02T11:14:01.494834Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-02T11:14:01.572570Z","title":null,"venue":null,"work_id":null,"year":2023},"citing_paper":{"arxiv_id":"2607.22614","last_updated":"2026-06-15T09:41:09Z","snapshot_observed_at":"2026-08-07T04:23:01.926858Z","submitted_at":"2026-06-15T09:41:09Z","title":"DynaResize: Runtime GPU Reallocation for Disaggregated LLM Post-Training","version":1},"reference_index":27,"source":"pdf_text","source_observed_at":"2026-08-02T11:14:01.572570Z"},"links":{"citing_paper":"/paper/2607.22614"},"observation_digest":"sha256:f3f348d11f906ec0fc4ec7d7dd0c2aac71a0d868bd5b78e9ff41b62498305f3b","observation_id":"4fc783c1-1122-4b43-b0dc-ae9333ccf17c","resolution":{"observed_at":"2026-08-02T11:14:01.572570Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"1909.08053","last_updated":"2020-03-13T23:45:18Z","snapshot_observed_at":"2026-07-06T02:11:23.670680Z","submitted_at":"2019-09-17T19:42:54Z","title":"Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism","version":4},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"1909.08053","snapshot_observed_at":"2026-08-02T11:14:01.686672Z","title":"2019.Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism","venue":null,"work_id":null,"year":2019},"citing_paper":{"arxiv_id":"2607.22614","last_updated":"2026-06-15T09:41:09Z","snapshot_observed_at":"2026-08-07T04:23:01.926858Z","submitted_at":"2026-06-15T09:41:09Z","title":"DynaResize: Runtime GPU Reallocation for Disaggregated LLM Post-Training","version":1},"reference_index":28,"source":"pdf_text","source_observed_at":"2026-08-02T11:14:01.686672Z"},"links":{"cited_paper":"/paper/1909.08053","citing_paper":"/paper/2607.22614"},"observation_digest":"sha256:77b3655245b2505c2d7519e238d42e6e8ca74f380af45ac93d8e3abac7df0f43","observation_id":"ac469568-a099-45e2-8356-24ee1eecb66e","resolution":{"observed_at":"2026-08-02T11:14:01.686672Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2501.12599","last_updated":"2025-06-03T02:14:54Z","snapshot_observed_at":"2026-07-06T02:11:23.670680Z","submitted_at":"2025-01-22T02:48:14Z","title":"Kimi k1.5: Scaling Reinforcement Learning with LLMs","version":4},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2501.12599","snapshot_observed_at":"2026-08-02T11:14:01.757024Z","title":"2025.Kimi k1.5: Scaling Reinforcement Learning with LLMs","venue":null,"work_id":null,"year":2025},"citing_paper":{"arxiv_id":"2607.22614","last_updated":"2026-06-15T09:41:09Z","snapshot_observed_at":"2026-08-07T04:23:01.926858Z","submitted_at":"2026-06-15T09:41:09Z","title":"DynaResize: Runtime GPU Reallocation for Disaggregated LLM Post-Training","version":1},"reference_index":29,"source":"pdf_text","source_observed_at":"2026-08-02T11:14:01.757024Z"},"links":{"cited_paper":"/paper/2501.12599","citing_paper":"/paper/2607.22614"},"observation_digest":"sha256:1c58a1b2e80455f679c9294646663905fa29f8258a6d6567c8094b5ae0c75fb9","observation_id":"c840c3b6-9bbd-424e-8df7-fd78f017bb59","resolution":{"observed_at":"2026-08-02T11:14:01.757024Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2505.09388","last_updated":"2025-05-14T13:41:34Z","snapshot_observed_at":"2026-07-06T02:11:23.670680Z","submitted_at":"2025-05-14T13:41:34Z","title":"Qwen3 Technical Report","version":1},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2505.09388","snapshot_observed_at":"2026-08-02T11:14:01.805158Z","title":"2025.Qwen3 Technical Report","venue":null,"work_id":null,"year":2025},"citing_paper":{"arxiv_id":"2607.22614","last_updated":"2026-06-15T09:41:09Z","snapshot_observed_at":"2026-08-07T04:23:01.926858Z","submitted_at":"2026-06-15T09:41:09Z","title":"DynaResize: Runtime GPU Reallocation for Disaggregated LLM Post-Training","version":1},"reference_index":30,"source":"pdf_text","source_observed_at":"2026-08-02T11:14:01.805158Z"},"links":{"cited_paper":"/paper/2505.09388","citing_paper":"/paper/2607.22614"},"observation_digest":"sha256:d6674cee9488217063cbfb5f45b88145d110e4dc1d23c531971b845c5b16c23a","observation_id":"5a773f85-f1fd-4ca5-9695-c49bdf381ce5","resolution":{"observed_at":"2026-08-02T11:14:01.805158Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-02T11:14:01.895237Z","title":null,"venue":null,"work_id":null,"year":2023},"citing_paper":{"arxiv_id":"2607.22614","last_updated":"2026-06-15T09:41:09Z","snapshot_observed_at":"2026-08-07T04:23:01.926858Z","submitted_at":"2026-06-15T09:41:09Z","title":"DynaResize: Runtime GPU Reallocation for Disaggregated LLM Post-Training","version":1},"reference_index":31,"source":"pdf_text","source_observed_at":"2026-08-02T11:14:01.895237Z"},"links":{"citing_paper":"/paper/2607.22614"},"observation_digest":"sha256:6ccdf03348c8f1016f7d1a76a4ea27e97516f0e11b8f02d156dee4d1a38c079b","observation_id":"386e4ac9-87be-4ee8-a8d3-70af9b5d3e3a","resolution":{"observed_at":"2026-08-02T11:14:01.895237Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-02T11:14:01.967587Z","title":null,"venue":null,"work_id":null,"year":2026},"citing_paper":{"arxiv_id":"2607.22614","last_updated":"2026-06-15T09:41:09Z","snapshot_observed_at":"2026-08-07T04:23:01.926858Z","submitted_at":"2026-06-15T09:41:09Z","title":"DynaResize: Runtime GPU Reallocation for Disaggregated LLM Post-Training","version":1},"reference_index":32,"source":"pdf_text","source_observed_at":"2026-08-02T11:14:01.967587Z"},"links":{"citing_paper":"/paper/2607.22614"},"observation_digest":"sha256:1b1c58a2709ae2886600d6abfb6c11cbfd5f173bf543a196834ba8efaf479d49","observation_id":"473eea03-8611-4288-b0c2-cd082772b9ec","resolution":{"observed_at":"2026-08-02T11:14:01.967587Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2602.02192","last_updated":"2026-05-26T02:43:34Z","snapshot_observed_at":"2026-08-09T04:10:49.998392Z","submitted_at":"2026-02-02T14:57:53Z","title":"ECHO-2: A Large-Scale Distributed Rollout Framework for Cost-Efficient Reinforcement Learning","version":5},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2602.02192","snapshot_observed_at":"2026-08-02T11:14:02.042304Z","title":"2026.ECHO-2: A Large-Scale Distributed Rollout Framework for Cost-Efficient Reinforcement Learning","venue":null,"work_id":null,"year":2026},"citing_paper":{"arxiv_id":"2607.22614","last_updated":"2026-06-15T09:41:09Z","snapshot_observed_at":"2026-08-07T04:23:01.926858Z","submitted_at":"2026-06-15T09:41:09Z","title":"DynaResize: Runtime GPU Reallocation for Disaggregated LLM Post-Training","version":1},"reference_index":33,"source":"pdf_text","source_observed_at":"2026-08-02T11:14:02.042304Z"},"links":{"cited_paper":"/paper/2602.02192","citing_paper":"/paper/2607.22614"},"observation_digest":"sha256:ecbf8619a8d4db231455dd4c4e324fefec1c7e205baeb31b8600cda0e1e6c2ef","observation_id":"73182606-3436-4471-b46d-db063113c303","resolution":{"observed_at":"2026-08-02T11:14:02.042304Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-02T11:14:02.148938Z","title":"2025.RLinf: Flexible and Efficient Large-scale Reinforcement Learning via Macro-to-Micro Flow Transformation","venue":null,"work_id":null,"year":2025},"citing_paper":{"arxiv_id":"2607.22614","last_updated":"2026-06-15T09:41:09Z","snapshot_observed_at":"2026-08-07T04:23:01.926858Z","submitted_at":"2026-06-15T09:41:09Z","title":"DynaResize: Runtime GPU Reallocation for Disaggregated LLM Post-Training","version":1},"reference_index":34,"source":"pdf_text","source_observed_at":"2026-08-02T11:14:02.148938Z"},"links":{"citing_paper":"/paper/2607.22614"},"observation_digest":"sha256:e94287277ed85f3a9765b60506cea21234c6e8a8dfdfb715d7933df652632960","observation_id":"2301aad0-1026-4c15-8f8b-bf7f81c3ee77","resolution":{"observed_at":"2026-08-02T11:14:02.148938Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-02T11:14:02.285564Z","title":null,"venue":null,"work_id":null,"year":2022},"citing_paper":{"arxiv_id":"2607.22614","last_updated":"2026-06-15T09:41:09Z","snapshot_observed_at":"2026-08-07T04:23:01.926858Z","submitted_at":"2026-06-15T09:41:09Z","title":"DynaResize: Runtime GPU Reallocation for Disaggregated LLM Post-Training","version":1},"reference_index":35,"source":"pdf_text","source_observed_at":"2026-08-02T11:14:02.285564Z"},"links":{"citing_paper":"/paper/2607.22614"},"observation_digest":"sha256:d088801750b72860a8636a67c30e87b46126f6cb801b0ca4da9a79f8a83a86c1","observation_id":"3a13db93-354b-424b-917e-237f01af6d8b","resolution":{"observed_at":"2026-08-02T11:14:02.285564Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2503.14476","last_updated":"2025-05-20T01:37:34Z","snapshot_observed_at":"2026-08-02T01:40:54.187278Z","submitted_at":"2025-03-18T17:49:06Z","title":"DAPO: An Open-Source LLM Reinforcement Learning System at Scale","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2503.14476","snapshot_observed_at":"2026-08-02T11:14:02.385685Z","title":"2025.DAPO: An Open-Source LLM Reinforcement Learning System at Scale","venue":null,"work_id":null,"year":2025},"citing_paper":{"arxiv_id":"2607.22614","last_updated":"2026-06-15T09:41:09Z","snapshot_observed_at":"2026-08-07T04:23:01.926858Z","submitted_at":"2026-06-15T09:41:09Z","title":"DynaResize: Runtime GPU Reallocation for Disaggregated LLM Post-Training","version":1},"reference_index":36,"source":"pdf_text","source_observed_at":"2026-08-02T11:14:02.385685Z"},"links":{"cited_paper":"/paper/2503.14476","citing_paper":"/paper/2607.22614"},"observation_digest":"sha256:b69498f08342d82670bdd2a15cfe80266b3923602e1ef4f7dd357ff74ccc474a","observation_id":"c81ed167-a5cd-4dd1-a8dc-b5e68d3bd6b0","resolution":{"observed_at":"2026-08-02T11:14:02.385685Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-02T11:14:02.528325Z","title":"2026.ProRL Agent: Rollout-as-a-Service for RL Training of Multi-Turn LLM Agents","venue":null,"work_id":null,"year":2026},"citing_paper":{"arxiv_id":"2607.22614","last_updated":"2026-06-15T09:41:09Z","snapshot_observed_at":"2026-08-07T04:23:01.926858Z","submitted_at":"2026-06-15T09:41:09Z","title":"DynaResize: Runtime GPU Reallocation for Disaggregated LLM Post-Training","version":1},"reference_index":37,"source":"pdf_text","source_observed_at":"2026-08-02T11:14:02.528325Z"},"links":{"citing_paper":"/paper/2607.22614"},"observation_digest":"sha256:b28cd6164dc00e0148333eb1b1f48538e98311f2b4946ccb9f9c585fe53e75c3","observation_id":"1c460363-5ad0-431e-a078-5c88ae885e9b","resolution":{"observed_at":"2026-08-02T11:14:02.528325Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-02T11:14:02.595933Z","title":null,"venue":null,"work_id":null,"year":2026},"citing_paper":{"arxiv_id":"2607.22614","last_updated":"2026-06-15T09:41:09Z","snapshot_observed_at":"2026-08-07T04:23:01.926858Z","submitted_at":"2026-06-15T09:41:09Z","title":"DynaResize: Runtime GPU Reallocation for Disaggregated LLM Post-Training","version":1},"reference_index":38,"source":"pdf_text","source_observed_at":"2026-08-02T11:14:02.595933Z"},"links":{"citing_paper":"/paper/2607.22614"},"observation_digest":"sha256:434894e785fa065d73df87e47293409cc4305a2d36e74fda37297b9cf8c3caf8","observation_id":"06e8427e-55ea-4cc8-aad2-8b4715de6660","resolution":{"observed_at":"2026-08-02T11:14:02.595933Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2304.11277","last_updated":"2023-09-12T16:28:00Z","snapshot_observed_at":"2026-08-01T19:01:47.393546Z","submitted_at":"2023-04-21T23:52:27Z","title":"PyTorch FSDP: Experiences on Scaling Fully Sharded Data Parallel","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2304.11277","snapshot_observed_at":"2026-08-02T11:14:02.713440Z","title":"2023.PyTorch FSDP: Experiences on Scaling Fully Sharded Data Parallel","venue":null,"work_id":null,"year":2023},"citing_paper":{"arxiv_id":"2607.22614","last_updated":"2026-06-15T09:41:09Z","snapshot_observed_at":"2026-08-07T04:23:01.926858Z","submitted_at":"2026-06-15T09:41:09Z","title":"DynaResize: Runtime GPU Reallocation for Disaggregated LLM Post-Training","version":1},"reference_index":39,"source":"pdf_text","source_observed_at":"2026-08-02T11:14:02.713440Z"},"links":{"cited_paper":"/paper/2304.11277","citing_paper":"/paper/2607.22614"},"observation_digest":"sha256:36ce468542882f7335cb259ef9e9d5dcc7627b4d62738a7282c92c8d9708a66f","observation_id":"53a5e31d-1727-43dd-8941-23f6ea389511","resolution":{"observed_at":"2026-08-02T11:14:02.713440Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-02T11:14:02.821325Z","title":"2025.APRIL: Active Partial Rollouts in Reinforcement Learning to tame long-tail generation","venue":null,"work_id":null,"year":2025},"citing_paper":{"arxiv_id":"2607.22614","last_updated":"2026-06-15T09:41:09Z","snapshot_observed_at":"2026-08-07T04:23:01.926858Z","submitted_at":"2026-06-15T09:41:09Z","title":"DynaResize: Runtime GPU Reallocation for Disaggregated LLM Post-Training","version":1},"reference_index":40,"source":"pdf_text","source_observed_at":"2026-08-02T11:14:02.821325Z"},"links":{"citing_paper":"/paper/2607.22614"},"observation_digest":"sha256:dfaa45c21bfb2b0c9bd970a9fea6e1b8018e1f0bd8afb3e2ea4cabcfb163c8a3","observation_id":"93385d83-f220-429d-b857-c316bd578466","resolution":{"observed_at":"2026-08-02T11:14:02.821325Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-02T11:14:02.893027Z","title":null,"venue":null,"work_id":null,"year":2025},"citing_paper":{"arxiv_id":"2607.22614","last_updated":"2026-06-15T09:41:09Z","snapshot_observed_at":"2026-08-07T04:23:01.926858Z","submitted_at":"2026-06-15T09:41:09Z","title":"DynaResize: Runtime GPU Reallocation for Disaggregated LLM Post-Training","version":1},"reference_index":41,"source":"pdf_text","source_observed_at":"2026-08-02T11:14:02.893027Z"},"links":{"citing_paper":"/paper/2607.22614"},"observation_digest":"sha256:bd8ef1bc511e605131b7c9404d055ddbd5f0fedbdf24558f408556867f97a4a3","observation_id":"614a081e-182b-4527-bd5f-f3134f7b24d5","resolution":{"observed_at":"2026-08-02T11:14:02.893027Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}}],"paper":{"arxiv_id":"2607.22614","last_updated":"2026-06-15T09:41:09Z","latest_version":1,"primary_category":"cs.AI","snapshot_observed_at":"2026-08-07T04:23:01.926858Z","submitted_at":"2026-06-15T09:41:09Z","title":"DynaResize: Runtime GPU Reallocation for Disaggregated LLM Post-Training"},"reference_resolution":{"displayed":41,"state_counts":{"malformed_identifier":0,"metadata_mismatch":0,"parse_uncertain":0,"unresolved":41,"verified_exact":0,"verified_fuzzy":0},"total_outbound_references":41},"refusal":"A citation records a reference. It does not transfer a finding from one paper to another.","schema":"pith.paper-citation-record.v1","standing_sources":[{"observed_at":"2026-08-09T06:31:02.800959+00:00","source":"crossref"},{"observed_at":"2026-08-09T06:30:57.326959+00:00","source":"retraction_watch"}],"thesis":"As of 9 August 2026, this Paper Citation Record lists 41 of 41 outbound references and 0 inbound Pith citation observations for arXiv:2607.22614."}