Introduces RSI metric and RSI-S filtering method for adaptive token selection in RLVR, reporting 2-3 point gains over GRPO on AIME/AMC benchmarks.
Fipo: Eliciting deep reasoning with future-kl influenced policy optimization
13 Pith papers cite this work. Polarity classification is still indexing.
13
Pith papers citing it
citation-role summary
background 1
baseline 1
citation-polarity summary
years
2026 13representative citing papers
PASS middleware independently standardizes process/outcome/format streams, derives value-homogeneous chunks, and converts cumulative returns to average value density, yielding consistent pass@1 gains over GRPO baselines in two domains and two signal paradigms.
ReSum's contrastive RL branching on self-summarization points improves LLM math reasoning accuracy by about 4% and shortens rollouts by about 18.6% across tested backbones.
CPPO replaces uniform token-level trust regions in PPO-style RLVR with position-weighted thresholds and cumulative prefix budgets to better align with autoregressive generation.
Optimal Coefficient Calibration (OCC) enables joint MTP-RL training to match or exceed the detach baseline on six math reasoning benchmarks by tracking the optimal coefficient online.
DelTA estimates token coefficients to amplify discriminative directions in token-gradient vectors, reweighting the RLVR surrogate to produce more contrastive side-wise centroids and yielding 3.26 and 2.62 point gains on math benchmarks for 8B and 14B Qwen3 models.
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.
ActFocus resolves the action bottleneck in agentic RL by reweighting token gradients toward action tokens using observed reward variance and an energy-based uncertainty term, outperforming PPO and GRPO by up to 65 percentage points.
DGPO is a critic-free RL framework that uses bounded Hellinger distance and entropy-gated advantage redistribution to enable fine-grained token-level credit assignment in long CoT generations for LLM alignment, reporting SOTA results on AIME benchmarks.
VIMPO derives a policy-implied value function from optimality conditions for critic-free RL in LLMs and shows gains over GRPO on math benchmarks.
APPO improves LLM agent training by branching at tokens selected for both uncertainty and future impact, then scaling credit for consequential reasoning procedures.
OWPO decouples optimization direction from magnitude via asymmetric reweighting (Accelerated Alignment for inferior deviations, Gain Locking for superior) plus iterative references to create a ratchet effect for continuous LLM improvement.
VISD proposes structured self-distillation with a multi-dimensional judge model and direction-magnitude decoupling to improve token-level credit assignment and convergence speed in VideoLLM reasoning training.
citing papers explorer
-
Which Tokens Matter? Adaptive Token Selection for RLVR with the Relative Surprisal Index
Introduces RSI metric and RSI-S filtering method for adaptive token selection in RLVR, reporting 2-3 point gains over GRPO on AIME/AMC benchmarks.
-
Process Advantage Signal Shaping: A Paradigm-Agnostic Middleware for Process-Supervised RL in LLM Reasoners
PASS middleware independently standardizes process/outcome/format streams, derives value-homogeneous chunks, and converts cumulative returns to average value density, yielding consistent pass@1 gains over GRPO baselines in two domains and two signal paradigms.
-
ReSum: Synergizing LLM Reasoning and Summarization with Reinforcement Learning
ReSum's contrastive RL branching on self-summarization points improves LLM math reasoning accuracy by about 4% and shortens rollouts by about 18.6% across tested backbones.
-
Beyond Uniform Token-Level Trust Region in LLM Reinforcement Learning
CPPO replaces uniform token-level trust regions in PPO-style RLVR with position-weighted thresholds and cumulative prefix budgets to better align with autoregressive generation.
-
Joint Training of Multi-Token Prediction in Reinforcement Learning via Optimal Coefficient Calibration
Optimal Coefficient Calibration (OCC) enables joint MTP-RL training to match or exceed the detach baseline on six math reasoning benchmarks by tracking the optimal coefficient online.
-
DelTA: Discriminative Token Credit Assignment for Reinforcement Learning from Verifiable Rewards
DelTA estimates token coefficients to amplify discriminative directions in token-gradient vectors, reweighting the RLVR surrogate to produce more contrastive side-wise centroids and yielding 3.26 and 2.62 point gains on math benchmarks for 8B and 14B Qwen3 models.
-
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.
-
Resolving Action Bottleneck: Agentic Reinforcement Learning Informed by Token-Level Energy
ActFocus resolves the action bottleneck in agentic RL by reweighting token gradients toward action tokens using observed reward variance and an energy-based uncertainty term, outperforming PPO and GRPO by up to 65 percentage points.
-
DGPO: Distribution Guided Policy Optimization for Fine Grained Credit Assignment
DGPO is a critic-free RL framework that uses bounded Hellinger distance and entropy-gated advantage redistribution to enable fine-grained token-level credit assignment in long CoT generations for LLM alignment, reporting SOTA results on AIME benchmarks.
-
VIMPO: Value-Implicit Policy Optimization for LLMs
VIMPO derives a policy-implied value function from optimality conditions for critic-free RL in LLMs and shows gains over GRPO on math benchmarks.
-
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.
-
One-Way Policy Optimization for Self-Evolving LLMs
OWPO decouples optimization direction from magnitude via asymmetric reweighting (Accelerated Alignment for inferior deviations, Gain Locking for superior) plus iterative references to create a ratchet effect for continuous LLM improvement.
-
VISD: Enhancing Video Reasoning via Structured Self-Distillation
VISD proposes structured self-distillation with a multi-dimensional judge model and direction-magnitude decoupling to improve token-level credit assignment and convergence speed in VideoLLM reasoning training.