ATLAS traces RLVR data to 20 atomic sources, most datasets are variants, and DAPO++ curated with SCA improves RLVR performance while Q predicts training effectiveness.
Klear-reasoner: Advancing reasoning capability via gradient-preserving clipping policy optimization.arXiv preprint arXiv:2508.07629
13 Pith papers cite this work, alongside 1 external citations. Polarity classification is still indexing.
13
Pith papers citing it
1
external citations · external index
citation-role summary
background 1
method 1
citation-polarity summary
polarities
background 2representative citing papers
Pion modifies Muon's Newton-Schulz iterations into a controllable high-pass filter that anchors dominant singular values at 1 while suppressing noisy tails, outperforming Muon and AdamW in VLA and RLVR regimes.
DuST self-trains LLMs for code generation by ranking their own test-time samples via sandbox execution and applying GRPO, improving judgment by +6.2 NDCG and single-sample pass@1 by +3.1 on LiveCodeBench.
Replacing the importance sampling ratio with a stop-gradient self-anchored ratio for positive advantages yields unclipped, REINFORCE-equivalent gradients that improve exploration without training instability.
Theoretical analysis of RLVR update dynamics leads to ACPO, an adaptive clipping method that outperforms DAPO and CISPO on reasoning benchmarks with 3B and 7B models.
D-VLA uses plane decoupling and a swimlane pipeline to deliver higher throughput and linear speedup than prior RL frameworks when training billion- and trillion-parameter VLA models on benchmarks like LIBERO.
Missing old logits in async agentic RL entangle discrepancy and staleness terms in PPO off-policy correction; exact acquisition methods and revised PPO-EWMA restore decoupled updates with reported gains in speed and performance.
Entropy Ratio Clipping introduces a global entropy-ratio constraint that stabilizes RL policy updates in LLM post-training beyond local PPO clipping.
CE-GPPO preserves bounded gradients from clipped tokens in PPO to regulate entropy evolution and improve performance on mathematical reasoning benchmarks.
Survey mapping RL techniques onto LLM training and highlighting gaps in value-based, off-policy, and bootstrapping methods.
APPO improves LLM agent training by branching at tokens selected for both uncertainty and future impact, then scaling credit for consequential reasoning procedures.
R²VPO uses ratio-variance regularization as a distributional soft brake on policy updates, claiming better performance than PPO on math reasoning and robotic control without hard clipping.
OneRec-V2 scales generative recommendation to 8B parameters via decoder-only design and real-world preference alignment, improving user engagement metrics in production A/B tests.
citing papers explorer
-
RLVR Datasets and Where to Find Them: Tracing Data Lineage for Better Training Data
ATLAS traces RLVR data to 20 atomic sources, most datasets are variants, and DAPO++ curated with SCA improves RLVR performance while Q predicts training effectiveness.
-
Rethinking Muon Beyond Pretraining: Spectral Failures and High-Pass Remedies for VLA and RLVR
Pion modifies Muon's Newton-Schulz iterations into a controllable high-pass filter that anchors dominant singular values at 1 while suppressing noisy tails, outperforming Muon and AdamW in VLA and RLVR regimes.
-
Primal Generation, Dual Judgment: Self-Training from Test-Time Scaling
DuST self-trains LLMs for code generation by ranking their own test-time samples via sandbox execution and applying GRPO, improving judgment by +6.2 NDCG and single-sample pass@1 by +3.1 on LiveCodeBench.
-
UP: Unbounded Positive Asymmetric Optimization for Breaking the Exploration-Stability Dilemma
Replacing the importance sampling ratio with a stop-gradient self-anchored ratio for positive advantages yields unclipped, REINFORCE-equivalent gradients that improve exploration without training instability.
-
What are Key Factors for Updates in RL for LLM Reasoning?
Theoretical analysis of RLVR update dynamics leads to ACPO, an adaptive clipping method that outperforms DAPO and CISPO on reasoning benchmarks with 3B and 7B models.
-
D-VLA: A High-Concurrency Distributed Asynchronous Reinforcement Learning Framework for Vision-Language-Action Models
D-VLA uses plane decoupling and a swimlane pipeline to deliver higher throughput and linear speedup than prior RL frameworks when training billion- and trillion-parameter VLA models on benchmarks like LIBERO.
-
Missing Old Logits in Asynchronous Agentic RL: Semantic Mismatch and Repair Methods for Off-Policy Correction
Missing old logits in async agentic RL entangle discrepancy and staleness terms in PPO off-policy correction; exact acquisition methods and revised PPO-EWMA restore decoupled updates with reported gains in speed and performance.
-
Entropy Ratio Clipping as a Soft Global Constraint for Stable Reinforcement Learning
Entropy Ratio Clipping introduces a global entropy-ratio constraint that stabilizes RL policy updates in LLM post-training beyond local PPO clipping.
-
CE-GPPO: Coordinating Entropy via Gradient-Preserving Clipping Policy Optimization in Reinforcement Learning
CE-GPPO preserves bounded gradients from clipped tokens in PPO to regulate entropy evolution and improve performance on mathematical reasoning benchmarks.
-
Modularized Reinforcement Learning on LLMs: From MDP Creation to Exploration and Learning
Survey mapping RL techniques onto LLM training and highlighting gaps in value-based, off-policy, and bootstrapping methods.
-
APPO: Agentic Procedural Policy Optimization
APPO improves LLM agent training by branching at tokens selected for both uncertainty and future impact, then scaling credit for consequential reasoning procedures.
-
Ratio-Variance Regularized Policy Optimization
R²VPO uses ratio-variance regularization as a distributional soft brake on policy updates, claiming better performance than PPO on math reasoning and robotic control without hard clipping.
-
OneRec-V2 Technical Report
OneRec-V2 scales generative recommendation to 8B parameters via decoder-only design and real-world preference alignment, improving user engagement metrics in production A/B tests.