StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training | Paper Citation Record

Observation 8b8c4b7e-93e3-48ba-b5b7-468fa1717476 · outbound

This paper cites Taming Throughput-Latency tradeoff in LLM inference with Sarathi-Serve.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Taming Throughput-Latency tradeoff in LLM inference with Sarathi-Serve

Reference 1

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:45.902248Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 907c093c-b1ae-4d7f-8fc0-47545e2e0262 · outbound

This paper cites LongAlign: A Recipe for Long Context Alignment of Large Language Models.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training LongAlign: A Recipe for Long Context Alignment of Large Language Models

Reference 2

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:46.082302Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 50e8d2b0-f64a-4992-97c6-e3d020b4a30e · outbound

This paper cites A survey on mixture of experts in large language models.IEEE Transactions on Knowledge and Data Engineering, 2025.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training A survey on mixture of experts in large language models.IEEE Transactions on Knowledge and Data Engineering, 2025

Reference 3

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:46.235143Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 5f8c24d9-c3b9-4544-8674-9d055a58ceda · outbound

This paper cites Respec: Towards optimizing speculative decoding in reinforcement learning systems, 2025.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Respec: Towards optimizing speculative decoding in reinforcement learning systems, 2025

Reference 4

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:46.402975Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation d972a8dc-1106-4639-b004-12dd92cb13a4 · outbound

This paper cites Deepseek-r1: Incentivizing reasoning capability in llms via reinforcement learning, 2025.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Deepseek-r1: Incentivizing reasoning capability in llms via reinforcement learning, 2025

Reference 5

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:46.529447Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation aa62aa5c-4434-43ff-90ea-d3e69ecc663c · outbound

This paper cites UloRL:An Ultra-Long Output Reinforcement Learning Approach for Advancing Large Language Models' Reasoning Abilities.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training UloRL:An Ultra-Long Output Reinforcement Learning Approach for Advancing Large Language Models' Reasoning Abilities

Reference 6

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:46.677712Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 225f3113-8c15-4110-8143-e0388d866bc8 · outbound

This paper cites Advances in importance sampling.Wiley Stat- sRef: Statistics Reference Online, 2021.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Advances in importance sampling.Wiley Stat- sRef: Statistics Reference Online, 2021

Reference 7

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:46.835994Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation fc8debaf-3b4e-4486-a0b3-32dd89500f20 · outbound

This paper cites Aime24 dataset, 2025.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Aime24 dataset, 2025

Reference 8

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:47.009776Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation a17c8c52-63a4-41c4-aa48-bff89400eafb · outbound

This paper cites Dapo-math-17k dataset, 2025.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Dapo-math-17k dataset, 2025

Reference 9

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:47.120583Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 7d2d5ef5-381b-41af-b18b-44f0ae5fc23f · outbound

This paper cites AReaL: A Large-Scale Asynchronous Reinforcement Learning System for Language Reasoning.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training AReaL: A Large-Scale Asynchronous Reinforcement Learning System for Language Reasoning

Reference 10

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:47.347729Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 38d80a14-fb98-42a9-96b4-219a77a70d4f · outbound

This paper cites Apt-serve: Adaptive request scheduling on hybrid cache for scalable llm inference serving.Proc.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Apt-serve: Adaptive request scheduling on hybrid cache for scalable llm inference serving.Proc

Reference 11

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:47.508488Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 6dc25872-f2f3-4214-a7fe-0dfde2e71a0a · outbound

This paper cites Rollpacker: Mitigating long-tail rollouts for fast, synchronous rl post-training, 2025.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Rollpacker: Mitigating long-tail rollouts for fast, synchronous rl post-training, 2025

Reference 12

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:47.514874Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 6a730f45-6758-40d8-a961-4c2314fea167 · outbound

This paper cites Enabling parallelism hot switching for efficient training of large language models.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Enabling parallelism hot switching for efficient training of large language models

Reference 13

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:47.520984Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 5eebe9cf-6b28-48fb-8597-d82a73133610 · outbound

This paper cites Search and score-based waterfall auction optimization.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Search and score-based waterfall auction optimization

Reference 14

Resolution: verified exact; doi, observed 2026-08-04T06:28:25.768871Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-23T06:30:58.430688+00:00.

Observation b50643ea-e6e5-4f20-be01-0a8656750e66 · outbound

This paper cites AsyncFlow: An Asynchronous Streaming RL Framework for Efficient LLM Post-Training.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training AsyncFlow: An Asynchronous Streaming RL Framework for Efficient LLM Post-Training

Reference 15

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:47.686864Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation a46dc5c9-5719-4b7b-b5d7-0787bff3f378 · outbound

This paper cites History Rhymes: Accelerating LLM Reinforcement Learning with RhymeRL.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training History Rhymes: Accelerating LLM Reinforcement Learning with RhymeRL

Reference 16

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:47.766365Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation cfdb1be1-1e9b-42bd-baf7-83a32845d875 · outbound

This paper cites Verl recipe: Fully async policy trainer, 2025.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Verl recipe: Fully async policy trainer, 2025

Reference 17

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:47.924312Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 1d2fe562-2015-4b76-b712-e7ba8f7aa53f · outbound

This paper cites Verl recipe: One step off policy async trainer,.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Verl recipe: One step off policy async trainer,

Reference 18

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:48.129022Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation e53b63b9-fc9c-4724-8b3f-55c1464a02da · outbound

This paper cites OpenRLHF: An Easy-to-use, Scalable and High-performance RLHF Framework.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training OpenRLHF: An Easy-to-use, Scalable and High-performance RLHF Framework

Reference 19

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:48.406504Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation eb9fcff0-e087-408a-94f2-21c73d46cd85 · outbound

This paper cites Demystifying nccl: An in-depth analysis of gpu communication protocols and algorithms, 2025.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Demystifying nccl: An in-depth analysis of gpu communication protocols and algorithms, 2025

Reference 20

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:48.539167Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 9941557c-3dd0-4160-addf-4dd5e4369839 · outbound

This paper cites Qerl: Beyond efficiency – quantization-enhanced reinforcement learning for llms, 2025.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Qerl: Beyond efficiency – quantization-enhanced reinforcement learning for llms, 2025

Reference 21

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:48.677239Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 0e7f2b6d-022e-405e-be6b-acbfa248bba4 · outbound

This paper cites Le, and Yonghui Chen.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Le, and Yonghui Chen

Reference 22

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:48.849801Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 1ad5ed8c-dd73-47b6-a113-3f07a8738e18 · outbound

This paper cites System optimizations for enabling training of extreme long sequence transformer models.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training System optimizations for enabling training of extreme long sequence transformer models

Reference 23

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:49.004334Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation b10af087-2781-4102-9693-97c692522776 · outbound

This paper cites Dynapipe: Optimizing multi-task training through dynamic pipelines.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Dynapipe: Optimizing multi-task training through dynamic pipelines

Reference 24

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:49.171345Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 60fdfe69-e7a5-42ec-929f-df86bb10a6d7 · outbound

This paper cites The Art of Scaling Reinforcement Learning Compute for LLMs.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training The Art of Scaling Reinforcement Learning Compute for LLMs

Reference 25

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:49.318391Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 759b0824-ebb4-4f4c-91be-fca8c99cb1c3 · outbound

This paper cites Waterfall Bandits: Learning to Sell Ads Online.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Waterfall Bandits: Learning to Sell Ads Online

Reference 26

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:49.508315Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 9ed6135a-d992-47f3-a2ab-9d8f18e04021 · outbound

This paper cites Efficient mem- ory management for large language model serving with pagedattention.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Efficient mem- ory management for large language model serving with pagedattention

Reference 27

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:49.682524Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation b6776c5a-93e5-464a-bd42-290c099a750b · outbound

This paper cites Puzzle: efficiently aligning large language models through light-weight context switch.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Puzzle: efficiently aligning large language models through light-weight context switch

Reference 28

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:49.806586Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 971c45f1-5afc-4bae-9e6d-b00f6c7c220c · outbound

This paper cites {GS}hard: Scaling giant models with conditional computation and automatic sharding.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training {GS}hard: Scaling giant models with conditional computation and automatic sharding

Reference 29

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:49.925465Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation bdefbd6a-bae6-493f-aeee-3cce72f33190 · outbound

This paper cites Fast inference from transform- ers via speculative decoding.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Fast inference from transform- ers via speculative decoding

Reference 30

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:50.012227Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation b38a508e-bf0e-42df-b69c-a927050b39e0 · outbound

This paper cites Hetu v2: A general and scalable deep learning system with hierarchical and heterogeneous single program multiple data annotations,.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Hetu v2: A general and scalable deep learning system with hierarchical and heterogeneous single program multiple data annotations,

Reference 31

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:50.100852Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 42e09a2b-fdf6-4be4-91ec-89f98a815f44 · outbound

This paper cites Malleus: Straggler-resilient hybrid parallel training of large-scale models via malleable data and model parallelization.Proc.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Malleus: Straggler-resilient hybrid parallel training of large-scale models via malleable data and model parallelization.Proc

Reference 32

Resolution: verified exact; doi, observed 2026-08-04T06:28:25.646537Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-23T06:30:58.430688+00:00.

Observation 24802f64-f13b-4f1c-8a60-73b766f0c820 · outbound

This paper cites Hydraulis: Balancing large transformer model training via co-designing parallel strategies and data assignment.Proc.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Hydraulis: Balancing large transformer model training via co-designing parallel strategies and data assignment.Proc

Reference 33

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:50.348565Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 4e5f30a5-0322-405c-a3e0-dd3bf5e9586a · outbound

This paper cites Hetu v2: A General and Scalable Deep Learning System with Hierarchical and Heterogeneous Single Program Multiple Data Annotations.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Hetu v2: A General and Scalable Deep Learning System with Hierarchical and Heterogeneous Single Program Multiple Data Annotations

Reference 34

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:50.166691Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 562a772b-9e7f-402d-a5e7-21576d326cad · outbound

This paper cites Let’s verify step by step.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Let’s verify step by step

Reference 35

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:50.509055Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 4c74e100-221f-47cc-a6b4-c7d9d0dcc07e · outbound

This paper cites Lobra: Multi-tenant fine-tuning over heterogeneous data.Proc.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Lobra: Multi-tenant fine-tuning over heterogeneous data.Proc

Reference 36

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:50.724092Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation f253949f-05d5-43be-a2bf-d406335d44d2 · outbound

This paper cites Hotprefix: Hotness-aware kv cache scheduling for efficient prefix sharing in llm inference systems.Proc.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Hotprefix: Hotness-aware kv cache scheduling for efficient prefix sharing in llm inference systems.Proc

Reference 37

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:50.415563Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation da905cd3-48b8-4494-8335-61b8e644890f · outbound

This paper cites Ringattention with blockwise trans- formers for near-infinite context.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Ringattention with blockwise trans- formers for near-infinite context

Reference 38

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:50.889465Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 650fdb5a-4bf4-43ef-8ddd-ce6476a26ee5 · outbound

This paper cites an unresolved cited work.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Unresolved cited work

Reference 39

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:50.606722Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 920e84a1-2abf-46e8-a3ce-4ba31dab7334 · outbound

This paper cites Flashrl: 8bit rollouts, full power rl, 2025.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Flashrl: 8bit rollouts, full power rl, 2025

Reference 40

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:51.048688Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 854ae4ba-1565-41a9-88c3-8bfcbaff2279 · outbound

This paper cites Spec-rl: Accelerating on-policy reinforce- ment learning with speculative rollouts, 2026.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Spec-rl: Accelerating on-policy reinforce- ment learning with speculative rollouts, 2026

Reference 41

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:50.785172Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 4562a217-4d99-4aab-ba01-2804c3aa3179 · outbound

This paper cites Deepcoder: A fully open-source 14b coder at o3- mini level, 2025.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Deepcoder: A fully open-source 14b coder at o3- mini level, 2025

Reference 42

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:51.207314Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 13681cab-9ead-46f5-9a41-d593f8293177 · outbound

This paper cites When speed kills stability: Demystifying rl collapse from the inference- training mismatch, 2025.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training When speed kills stability: Demystifying rl collapse from the inference- training mismatch, 2025

Reference 43

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:50.976916Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 8f33d82f-ada1-4c8e-a12b-351bb4ddaf7a · outbound

This paper cites Galvatron: Efficient transformer training over multiple gpus using automatic parallelism.Proc.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Galvatron: Efficient transformer training over multiple gpus using automatic parallelism.Proc

Reference 44

Resolution: verified exact; doi, observed 2026-08-04T06:28:25.468395Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-23T06:30:58.430688+00:00.

Observation d6481e17-341b-4ad4-8d04-d40f5bc6e55f · outbound

This paper cites Part ii: Roll flash – accelerating rlvr and agentic training with asynchrony, 2025.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Part ii: Roll flash – accelerating rlvr and agentic training with asynchrony, 2025

Reference 45

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:51.130997Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 96516336-859e-4d2d-9fca-43c12dca1d7b · outbound

This paper cites Devanur, Gregory R.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Devanur, Gregory R

Reference 46

Resolution: malformed identifier; no resolver link, observed 2026-08-04T06:25:51.491354Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 82e09f00-0196-4482-8a9a-f59041cd9a3a · outbound

This paper cites Real: Efficient RLHF training of large language models with parameter reallocation.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Real: Efficient RLHF training of large language models with parameter reallocation

Reference 47

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:51.286478Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 29d17927-1ebc-4160-a6e4-eacac569bc0b · outbound

This paper cites Asynchronous RLHF: Faster and More Efficient Off-Policy RL for Language Models.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Asynchronous RLHF: Faster and More Efficient Off-Policy RL for Language Models

Reference 48

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:51.694220Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 412432b0-32ad-419b-b4b4-7f8657f4e42d · outbound

This paper cites A comprehensive survey of mixture-of-experts: Algo- rithms, theory, and applications, 2025.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training A comprehensive survey of mixture-of-experts: Algo- rithms, theory, and applications, 2025

Reference 49

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:51.422548Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation cbcafe16-774f-4030-a824-b49b56b8e2e4 · outbound

This paper cites Nvidia inference xfer library (nixl), 2025.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Nvidia inference xfer library (nixl), 2025

Reference 50

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:51.842717Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation cd67b4a5-c2d9-4d92-8846-b060a70a52be · outbound

This paper cites Effi- cient large-scale language model training on gpu clusters using megatron-lm.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Effi- cient large-scale language model training on gpu clusters using megatron-lm

Reference 51

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:51.562805Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 3e190654-6ff3-499f-9558-835da6982878 · outbound

This paper cites Unified communication x, 2025.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Unified communication x, 2025

Reference 52

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:52.126406Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 4c1d3651-d1eb-47b4-b34e-246125ea5d41 · outbound

This paper cites Nvidia collective communication library (nccl) documentation, 2025.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Nvidia collective communication library (nccl) documentation, 2025

Reference 53

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:51.767667Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 2db48e94-488e-49d9-9b9f-5b2deaf7f653 · outbound

This paper cites Seer: Online Context Learning for Fast Synchronous LLM Reinforcement Learning.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Seer: Online Context Learning for Fast Synchronous LLM Reinforcement Learning

Reference 54

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:52.278491Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation eddd806c-6695-4a0f-96ee-c77cc7de796b · outbound

This paper cites Openai o1 system card,.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Openai o1 system card,

Reference 55

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:51.950676Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation cac33af7-6f70-45f9-936c-7400690331c7 · outbound

This paper cites OpenAI o1 System Card.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training OpenAI o1 System Card

Reference 56

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:52.052018Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation f84817e8-d325-4fd4-82d0-f84b5d17d5c5 · outbound

This paper cites Rosberg and I.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Rosberg and I

Reference 57

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:52.582254Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 9b11963d-20af-4e91-87d2-9c703e6a8a05 · outbound

This paper cites Multi-step reasoning with large language models, a survey.ACM Comput.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Multi-step reasoning with large language models, a survey.ACM Comput

Reference 58

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:52.209334Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 84380506-18cc-4997-8e56-5b6a75f96676 · outbound

This paper cites Proximal Policy Optimization Algorithms.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Proximal Policy Optimization Algorithms

Reference 59

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:52.769791Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation b0dba959-26dd-4795-9aba-fde77b0fd286 · outbound

This paper cites Qwen2.5 technical report,.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Qwen2.5 technical report,

Reference 60

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:52.382715Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation b00ef343-ed7e-4553-bf8f-68de283ebfeb · outbound

This paper cites Qwen2.5 Technical Report.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Qwen2.5 Technical Report

Reference 61

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:52.466334Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 61360b29-568e-4bbf-83af-e58566a5617d · outbound

This paper cites ZeRO: Memory Optimizations Toward Training Trillion Parameter Models.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training ZeRO: Memory Optimizations Toward Training Trillion Parameter Models

Reference 62

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:52.512470Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 0c014a9e-22d8-49ff-9bbb-972a2e8c9436 · outbound

This paper cites Hybridflow: A flexible and efficient rlhf framework.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Hybridflow: A flexible and efficient rlhf framework

Reference 63

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:53.226125Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 6ff19108-eee3-4297-a4dd-95b62144ea25 · outbound

This paper cites Trust region policy optimization.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Trust region policy optimization

Reference 64

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:52.665305Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 79d20292-6995-4f4a-918c-5155b8232b6f · outbound

This paper cites Scaling Behaviors of LLM Reinforcement Learning Post-Training: An Empirical Study in Mathematical Reasoning.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Scaling Behaviors of LLM Reinforcement Learning Post-Training: An Empirical Study in Mathematical Reasoning

Reference 65

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:53.415049Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation e695934d-434f-4b28-a7b5-a0305e3a12f5 · outbound

This paper cites Beat the long tail: Distribution-aware speculative decoding for rl training, 2025.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Beat the long tail: Distribution-aware speculative decoding for rl training, 2025

Reference 66

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:52.897121Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 955ac5a8-e82a-44c4-b9c0-288a9f769fd6 · outbound

This paper cites DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models

Reference 67

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:52.994986Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 07b5031f-76a9-4ff2-908e-fb6a77d7e4ae · outbound

This paper cites Laminar: A scalable asynchronous rl post-training framework, 2025.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Laminar: A scalable asynchronous rl post-training framework, 2025

Reference 68

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:53.075946Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation d97d48ad-1e62-438c-90d9-480af97718d7 · outbound

This paper cites A survey on large language models for mathematical reasoning.ACM Comput.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training A survey on large language models for mathematical reasoning.ACM Comput

Reference 69

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:53.844641Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 4612ab4b-409e-436f-918f-638a801a6d25 · outbound

This paper cites Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism

Reference 70

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:53.315719Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 50fb9431-bd93-4066-b33a-9e69d465bb74 · outbound

This paper cites Improving automatic parallel training via balanced memory workload optimization.IEEE Transactions on Knowledge and Data Engineering, August 2024.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Improving automatic parallel training via balanced memory workload optimization.IEEE Transactions on Knowledge and Data Engineering, August 2024

Reference 71

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:54.057071Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 28c2d01e-1ce4-4c4e-b6e8-1786e2da8467 · outbound

This paper cites Kimi K2: Open Agentic Intelligence.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Kimi K2: Open Agentic Intelligence

Reference 72

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:53.541884Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation d5bdd606-2987-4e21-b619-42fbc18bb306 · outbound

This paper cites Kimi k1.5: Scaling Reinforcement Learning with LLMs.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Kimi k1.5: Scaling Reinforcement Learning with LLMs

Reference 73

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:53.654690Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 17f4edf4-c35e-4fa6-8617-3f4c431f76c1 · outbound

This paper cites Tokdar and Robert E.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Tokdar and Robert E

Reference 74

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:53.746745Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 2ad77e30-4bf1-42a9-8208-ebeed40cd770 · outbound

This paper cites Loongserve: Efficiently serving long-context large language models with elastic sequence parallelism.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Loongserve: Efficiently serving long-context large language models with elastic sequence parallelism

Reference 75

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:54.456232Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 1acb3228-e099-4eb7-9fee-cabf3f8bd5a7 · outbound

This paper cites Reinforcement Learning Optimization for Large-Scale Learning: An Efficient and User-Friendly Scaling Library.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Reinforcement Learning Optimization for Large-Scale Learning: An Efficient and User-Friendly Scaling Library

Reference 76

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:53.985066Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 1ea59b38-0d3e-4802-8a51-36cf7a286158 · outbound

This paper cites An Adaptive Placement and Parallelism Framework for Accelerating RLHF Training.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training An Adaptive Placement and Parallelism Framework for Accelerating RLHF Training

Reference 77

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:54.634282Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 2a1928ea-f039-40a9-85ac-a0841bbe5f13 · outbound

This paper cites Flexsp: Accelerating large language model training via flexible sequence parallelism.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Flexsp: Accelerating large language model training via flexible sequence parallelism

Reference 78

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:54.151703Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 0fca62d3-9fac-4363-924b-df5ea3a77cc7 · outbound

This paper cites Reinforcement Learning with Verifiable Rewards Implicitly Incentivizes Correct Reasoning in Base LLMs.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Reinforcement Learning with Verifiable Rewards Implicitly Incentivizes Correct Reasoning in Base LLMs

Reference 79

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:54.268990Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation e4c14c07-922f-412e-ac39-730f9a9a07cc · outbound

This paper cites The multiqueue: A simple and fast relaxed concurrent priority queue.ACM Trans.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training The multiqueue: A simple and fast relaxed concurrent priority queue.ACM Trans

Reference 80

Resolution: verified exact; doi, observed 2026-08-04T06:28:25.315086Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-23T06:30:58.430688+00:00.

Observation eec37935-3da4-49bb-a1f5-921389fd9f31 · outbound

This paper cites Flashinfer: Efficient and customizable attention engine for LLM inference serving.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Flashinfer: Efficient and customizable attention engine for LLM inference serving

Reference 81

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:54.987759Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 4e20ecf2-a14a-4eec-9edc-ab8728858d75 · outbound

This paper cites LlamaRL: A Distributed Asynchronous Reinforcement Learning Framework for Efficient Large-scale LLM Training.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training LlamaRL: A Distributed Asynchronous Reinforcement Learning Framework for Efficient Large-scale LLM Training

Reference 82

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:54.516468Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 5835fd35-7ffa-4c75-9c17-5d29d42d2d18 · outbound

This paper cites Does reinforcement learning really incentivize reasoning capacity in LLMs beyond the base model? In2nd AI for Math Workshop @ ICML 2025, 2025.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Does reinforcement learning really incentivize reasoning capacity in LLMs beyond the base model? In2nd AI for Math Workshop @ ICML 2025, 2025

Reference 83

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:55.133042Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 02adb91b-d507-43bf-aba5-8937b5c69b74 · outbound

This paper cites Qwen3 Technical Report.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Qwen3 Technical Report

Reference 84

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:54.718668Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 4fbb9027-e47f-40b2-bf72-933a041c3656 · outbound

This paper cites Survey on knowledge distillation for large language models: Methods, evaluation, and application.ACM Trans.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Survey on knowledge distillation for large language models: Methods, evaluation, and application.ACM Trans

Reference 85

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:54.811385Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 78a1a2b6-cff0-4e93-9370-d5d938d93c0c · outbound

This paper cites Your efficient rl framework secretly brings you off-policy rl training, 2025.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Your efficient rl framework secretly brings you off-policy rl training, 2025

Reference 86

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:54.901122Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 6811364e-e69c-4297-bf0e-ee924a2a275b · outbound

This paper cites Small leak can sink a great ship–boost rl training on moe with icepop!, 2025.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Small leak can sink a great ship–boost rl training on moe with icepop!, 2025

Reference 87

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:55.483112Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation bdd10f48-5fa4-4737-a8bd-bbd3f600476f · outbound

This paper cites DAPO: An Open-Source LLM Reinforcement Learning System at Scale.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training DAPO: An Open-Source LLM Reinforcement Learning System at Scale

Reference 88

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:55.044439Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 425d4484-cecb-4b70-9866-6173253b653c · outbound

This paper cites Stabilizing reinforcement learning with llms: Formulation and practices, 2025.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Stabilizing reinforcement learning with llms: Formulation and practices, 2025

Reference 89

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:55.712023Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 243bd583-69b6-4d5b-b20a-d6b730dd79ba · outbound

This paper cites Pqcache: Product quantization-based kvcache for long context llm inference.Proc.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Pqcache: Product quantization-based kvcache for long context llm inference.Proc

Reference 90

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:55.249282Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation b0f89fda-f521-49ad-a589-cecfaddff8b7 · outbound

This paper cites A Survey of Reinforcement Learning for Large Reasoning Models.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training A Survey of Reinforcement Learning for Large Reasoning Models

Reference 91

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:55.329841Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 9bfca4bf-a9f5-4f54-8a30-aea5556b1d13 · outbound

This paper cites SortedRL: Accelerating RL training for LLMs through online length-aware scheduling.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training SortedRL: Accelerating RL training for LLMs through online length-aware scheduling

Reference 92

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:55.423450Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 8449252e-88cc-4f8c-a898-5018835a4ac2 · outbound

This paper cites Distserve: disaggregating prefill and decoding for goodput-optimized large language model serving.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Distserve: disaggregating prefill and decoding for goodput-optimized large language model serving

Reference 93

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:56.071060Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 38a0cd2c-1b2a-4242-823b-8d1da3555989 · outbound

This paper cites Pytorch fsdp: Experiences on scaling fully sharded data parallel.Proc.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Pytorch fsdp: Experiences on scaling fully sharded data parallel.Proc

Reference 94

Resolution: verified exact; doi, observed 2026-08-04T06:28:25.140551Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-23T06:30:58.430688+00:00.

Observation c447d8ad-1637-49c6-8bfc-04edd57a3944 · outbound

This paper cites Optimizing rlhf training for large language models with stage fusion.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Optimizing rlhf training for large language models with stage fusion

Reference 95

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:56.265063Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 5f7de689-6ae7-496a-a0ec-3fba33b4d0e7 · outbound

This paper cites Group Sequence Policy Optimization.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Group Sequence Policy Optimization

Reference 96

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:55.821502Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation e76a52d9-9bc9-4be7-b84f-d0dd6b5f1783 · outbound

This paper cites Prosperity before collapse: How far can off-policy rl reach with stale data on llms?, 2025.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Prosperity before collapse: How far can off-policy rl reach with stale data on llms?, 2025

Reference 97

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:55.909032Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 5fc58b4c-c25b-4b5f-90bd-1a1a04a51e7e · outbound

This paper cites Gonzalez, Clark Barrett, and Ying Sheng.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training Gonzalez, Clark Barrett, and Ying Sheng

Reference 98

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:55.987217Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 940e6faa-2206-42ea-b86d-21bf3ee3ce7c · outbound

This paper cites StreamRL: Scalable, Heterogeneous, and Elastic RL for LLMs with Disaggregated Stream Generation.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training StreamRL: Scalable, Heterogeneous, and Elastic RL for LLMs with Disaggregated Stream Generation

Reference 100

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:56.161461Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 227b2cd2-da42-4d57-82fe-30fc210e4840 · outbound

This paper cites April: Active partial rollouts in reinforcement learning to tame long-tail generation,.

StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training April: Active partial rollouts in reinforcement learning to tame long-tail generation,

Reference 102

Resolution: unresolved; no resolver link, observed 2026-08-04T06:25:56.310135Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Coverage vector

Reference resolution

Outbound references

Pith citing papers