Reflect-R1 introduces the first evidence-driven self-correction framework for long video understanding using a three-stage pipeline, stage-decoupled RL via SD-GRPO, and a 120K dataset to achieve SOTA on VideoMME and LongVideoBench.
Stop summation: Min-form credit assignment is all process reward model needs for reasoning
12 Pith papers cite this work. Polarity classification is still indexing.
12
Pith papers citing it
citation-role summary
background 3
baseline 1
citation-polarity summary
representative citing papers
STRACE improves long-horizon agent optimization by constructing a dependency graph to filter redundant traces and localize true root-cause modules for targeted policy updates.
Pairing outcome rewards with verifiable per-action path penalties reduces constraint violations nearly sixfold at equal task success, while a progress potential accelerates learning only where partial progress is reachable.
BALTO projects claim-level verification into balanced token-level rewards for RL-based hallucination mitigation in LLMs.
CASPO trains LLMs via iterative direct preference optimization so that token-level confidence tracks step-wise correctness, then applies Confidence-aware Thought pruning at inference to improve both reliability and speed on reasoning benchmarks.
Span-level Wasserstein distances between hidden-state distributions of correct and incorrect rollouts provide a self-supervised signal to reweight advantages in GRPO, improving fine-grained credit assignment on math and code tasks.
Geoalign curates rollouts in online LLM RL by learning a projector on hidden states to flag and replace directionally inconsistent examples, yielding higher final performance and less oscillation than baselines on dialogue and math tasks.
Survey mapping RL techniques onto LLM training and highlighting gaps in value-based, off-policy, and bootstrapping methods.
TrOPD stabilizes on-policy distillation for LLMs with trust-region learning, outlier estimation, and off-policy guidance, outperforming prior OPD methods on reasoning and code benchmarks.
Reshaping outcome rewards, process signals, and rollout comparability in GRPO raises strict compile-and-semantic accuracy in agentic code repair from 0.385 to 0.535 under weak feedback.
LoRR augments preference optimization methods like DPO with high-replay training, periodic resets to initial data/policy, and a hybrid objective to improve sample efficiency and reduce primacy bias on math and reasoning tasks.
A survey compiling RL methods, challenges, data resources, and applications for enhancing reasoning in large language models and large reasoning models since DeepSeek-R1.
citing papers explorer
-
Reflect-R1: Evidence-Driven Reflection for Self-Correction in Long Video Understanding
Reflect-R1 introduces the first evidence-driven self-correction framework for long video understanding using a three-stage pipeline, stage-decoupled RL via SD-GRPO, and a 120K dataset to achieve SOTA on VideoMME and LongVideoBench.
-
From Noisy Traces to Root Causes: Structural Trajectory Analysis and Causal Extraction for Agent Optimization
STRACE improves long-horizon agent optimization by constructing a dependency graph to filter redundant traces and localize true root-cause modules for targeted policy updates.
-
RLVP: Penalize the Path, Reward the Outcome
Pairing outcome rewards with verifiable per-action path penalties reduces constraint violations nearly sixfold at equal task success, while a progress potential accelerates learning only where partial progress is reachable.
-
BALTO: Balanced Token-Level Policy Optimization for Hallucination Mitigation
BALTO projects claim-level verification into balanced token-level rewards for RL-based hallucination mitigation in LLMs.
-
Confidence-Aware Alignment Makes Reasoning LLMs More Reliable
CASPO trains LLMs via iterative direct preference optimization so that token-level confidence tracks step-wise correctness, then applies Confidence-aware Thought pruning at inference to improve both reliability and speed on reasoning benchmarks.
-
Hidden States Know Where Reasoning Diverges: Credit Assignment via Span-Level Wasserstein Distance
Span-level Wasserstein distances between hidden-state distributions of correct and incorrect rollouts provide a self-supervised signal to reweight advantages in GRPO, improving fine-grained credit assignment on math and code tasks.
-
GEOALIGN: Geometric Rollout Curation for Robust LLM Reinforcement Learning
Geoalign curates rollouts in online LLM RL by learning a projector on hidden states to flag and replace directionally inconsistent examples, yielding higher final performance and less oscillation than baselines on dialogue and math tasks.
-
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.
-
Trust Region On-Policy Distillation
TrOPD stabilizes on-policy distillation for LLMs with trust-region learning, outlier estimation, and off-policy guidance, outperforming prior OPD methods on reasoning and code benchmarks.
-
Signal Reshaping for GRPO in Weak-Feedback Agentic Code Repair
Reshaping outcome rewards, process signals, and rollout comparability in GRPO raises strict compile-and-semantic accuracy in agentic code repair from 0.385 to 0.535 under weak feedback.
-
Sample-efficient LLM Optimization with Reset Replay
LoRR augments preference optimization methods like DPO with high-replay training, periodic resets to initial data/policy, and a hybrid objective to improve sample efficiency and reduce primacy bias on math and reasoning tasks.
-
A Survey of Reinforcement Learning for Large Reasoning Models
A survey compiling RL methods, challenges, data resources, and applications for enhancing reasoning in large language models and large reasoning models since DeepSeek-R1.