10 papers
Experience Augmented Policy Optimization for LLM Reasoning
Jinda Lu, Kexin Huang, Junkang Wu +7
Reinforcement Learning with Verifiable Rewards (RLVR) is a powerful paradigm for improving the reasoning capabilities of large language models (LLMs). However, existing RLVR method…
ARMOR: Stabilizing On-Policy LLM RL with Off-Policy Anchor Samples
Kexin Huang, Junkang Wu, Jinda Lu +7
Reinforcement learning (RL) has significantly enhanced the reasoning capabilities of large language models (LLMs), yet the training process remains notoriously fragile. In this wor…
Clipping Bottleneck: Stabilizing RLVR via Stochastic Recovery of Near-Boundary Signals
Shuo Yang, Jinda Lu, Chiyu Ma +8
Reinforcement Learning with Verifiable Rewards (RLVR) has emerged as a central paradigm for scaling LLM reasoning, yet its optimization often suffers from training instability and…
One-Way Policy Optimization for Self-Evolving LLMs
Shuo Yang, Jinda Lu, Kexin Huang +6
Reinforcement Learning with Verifiable Rewards (RLVR) has become a promising paradigm for scaling reasoning capabilities of Large Language Models (LLMs). However, the sparsity of b…
FIPO: Eliciting Deep Reasoning with Future-KL Influenced Policy Optimization
Chiyu Ma, Shuo Yang, Kexin Huang +7
We present Future-KL Influenced Policy Optimization (FIPO), a reinforcement learning algorithm designed to overcome reasoning bottlenecks in large language models. While GRPO style…
Bridging Perception and Reasoning: Token Reweighting for RLVR in Multimodal LLMs
Jinda Lu, Junkang Wu, Jinghan Li +6
Extending Reinforcement Learning with Verifiable Rewards (RLVR) to multimodal large language models (MLLMs) faces a fundamental challenge: their responses inherently interleave per…
Sparse but Critical: A Token-Level Analysis of Distributional Shifts in RLVR Fine-Tuning of LLMs
Haoming Meng, Kexin Huang, Shaohang Wei +6
Reinforcement learning with verifiable rewards (RLVR) has significantly improved reasoning in large language models (LLMs), yet the token-level mechanisms underlying these improvem…
On the Direction of RLVR Updates for LLM Reasoning: Identification and Exploitation
Kexin Huang, Haoming Meng, Junkang Wu +10
Reinforcement learning with verifiable rewards (RLVR) has substantially improved the reasoning capabilities of large language models. While existing analyses identify that RLVR-ind…
P-Aligner: Enabling Pre-Alignment of Language Models via Principled Instruction Synthesis
Feifan Song, Bofei Gao, Yifan Song +6
Large Language Models (LLMs) are expected to produce safe, helpful, and honest content during interaction with human users, but they frequently fail to align with such values when…
Well Begun is Half Done: Low-resource Preference Alignment by Weak-to-Strong Decoding
Feifan Song, Shaohang Wei, Wen Luo +4
Large Language Models (LLMs) require alignment with human preferences to avoid generating offensive, false, or meaningless content. Recently, low-resource methods for LLM alignment…