Self-distillation token rewards measure input-response-feedback pointwise mutual information, and CREDIT extracts the input-specific component with contrastive baselines to improve LLM reasoning performance.
hub Mixed citations
Hybridflow: A flexible and efficient rlhf framework
Mixed citation behavior. Most common role is method (38%).
15
Pith papers citing it
Method
38% of classified citations
hub tools
citation-role summary
method 4
baseline 2
other 2
citation-polarity summary
years
2026 15representative citing papers
GCPO uses team-level credit assignment via determinant volume over reward-weighted semantic embeddings to promote non-redundant correct reasoning paths, improving both accuracy and diversity in LLM training.
AgentForesight introduces an online auditor model that predicts decisive errors in multi-agent trajectories at the earliest step using a coarse-to-fine reinforcement learning recipe on a new curated dataset AFTraj-2K.
PropGuard is a propagation-aware framework for LLM-MAS that constructs dual-view spatio-temporal graphs, employs a GE-GRPO inspector to recover suspicious subgraphs, and applies source-guided remediation to lower attack success while preserving task performance.
OPPO derives token-level advantages for LLM RL via Bayesian recursion on oracle signals, recovering prior distillation methods as a special case and showing gains on math and code benchmarks.
NFPO augments the PPO surrogate with N-step forward traces to bridge local approximations and exact policy gradients, delivering tighter policy-improvement bounds and improved results on reasoning benchmarks.
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.
Anti-Self-Distillation reverses self-distillation signals via PMI to fix overconfidence on structural tokens, matching GRPO baseline accuracy 2-10x faster with up to 11.5 point gains across 4B-30B models.
ICR creates a virtual shorter distribution from shortest correct on-policy responses to regularize RL post-training toward concise yet accurate reasoning, improving the accuracy-length Pareto frontier on math and knowledge benchmarks.
LaST-R1 introduces a RL post-training method called LAPO that optimizes latent Chain-of-Thought reasoning in vision-language-action models, yielding 99.9% success on LIBERO and up to 22.5% real-world gains.
CodePivot uses Python as a pivot language plus an Aggressive-Partial-Functional RL reward to train a 7B model that outperforms much larger LLMs on multilingual code transpilation without parallel corpora.
CODA uses rollout-based difficulty signals to drive two gates that penalize verbosity on easy instances and promote deliberation on hard ones, cutting token use over 60% on simple tasks while maintaining accuracy.
RLVR exhibits implicit reward overfitting to training data and optimizes heavy-tailed singular spectra with rank-1 focus on reasoning capability.
RCoT-Seg uses GRPO-reinforced keyframe selection from a CoT-start corpus followed by SAM2 mask propagation to improve video object segmentation under implicit temporal instructions over prior MLLM sampling methods.
citing papers explorer
-
From Generic Correlation to Input-Specific Credit in On-Policy Self Distillation
Self-distillation token rewards measure input-response-feedback pointwise mutual information, and CREDIT extracts the input-specific component with contrastive baselines to improve LLM reasoning performance.
-
Breaking $\textit{Winner-Takes-All}$: Cooperative Policy Optimization Improves Diverse LLM Reasoning
GCPO uses team-level credit assignment via determinant volume over reward-weighted semantic embeddings to promote non-redundant correct reasoning paths, improving both accuracy and diversity in LLM training.
-
AgentForesight: Online Auditing for Early Failure Prediction in Multi-Agent Systems
AgentForesight introduces an online auditor model that predicts decisive errors in multi-agent trajectories at the earliest step using a coarse-to-fine reinforcement learning recipe on a new curated dataset AFTraj-2K.
-
PropGuard: Safeguarding LLM-MAS via Propagation-Aware Exploration and Remediation
PropGuard is a propagation-aware framework for LLM-MAS that constructs dual-view spatio-temporal graphs, employs a GE-GRPO inspector to recover suspicious subgraphs, and applies source-guided remediation to lower attack success while preserving task performance.
-
OPPO: Bayesian Value Recursion for Token-Level Credit Assignment in LLM Reasoning
OPPO derives token-level advantages for LLM RL via Bayesian recursion on oracle signals, recovering prior distillation methods as a special case and showing gains on math and code benchmarks.
-
Multi-Step Likelihood-Ratio Correction for Reinforcement Learning with Verifiable Rewards
NFPO augments the PPO surrogate with N-step forward traces to bridge local approximations and exact policy gradients, delivering tighter policy-improvement bounds and improved results on reasoning benchmarks.
-
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.
-
Anti-Self-Distillation for Reasoning RL via Pointwise Mutual Information
Anti-Self-Distillation reverses self-distillation signals via PMI to fix overconfidence on structural tokens, matching GRPO baseline accuracy 2-10x faster with up to 11.5 point gains across 4B-30B models.
-
Implicit Compression Regularization: Concise Reasoning via Internal Shorter Distributions in RL Post-Training
ICR creates a virtual shorter distribution from shortest correct on-policy responses to regularize RL post-training toward concise yet accurate reasoning, improving the accuracy-length Pareto frontier on math and knowledge benchmarks.
-
LaST-R1: Reinforcing Robotic Manipulation via Adaptive Physical Latent Reasoning
LaST-R1 introduces a RL post-training method called LAPO that optimizes latent Chain-of-Thought reasoning in vision-language-action models, yielding 99.9% success on LIBERO and up to 22.5% real-world gains.
-
CodePivot: Bootstrapping Multilingual Transpilation in LLMs via Reinforcement Learning without Parallel Corpora
CodePivot uses Python as a pivot language plus an Aggressive-Partial-Functional RL reward to train a 7B model that outperforms much larger LLMs on multilingual code transpilation without parallel corpora.
-
CODA: Difficulty-Aware Compute Allocation for Adaptive Reasoning
CODA uses rollout-based difficulty signals to drive two gates that penalize verbosity on easy instances and promote deliberation on hard ones, cutting token use over 60% on simple tasks while maintaining accuracy.
-
On the Implicit Reward Overfitting and the Low-rank Dynamics in RLVR
RLVR exhibits implicit reward overfitting to training data and optimizes heavy-tailed singular spectra with rank-1 focus on reasoning capability.
-
RCoT-Seg: Reinforced Chain-of-Thought for Video Reasoning and Segmentation
RCoT-Seg uses GRPO-reinforced keyframe selection from a CoT-start corpus followed by SAM2 mask propagation to improve video object segmentation under implicit temporal instructions over prior MLLM sampling methods.