11 papers
CVPO: Enhancing LLM Reinforcement Learning Reasoning via Value-Variance Adaptation and Dynamic Curriculum Learning
Ziqi Jia, Yalu Ouyang, Bo Pang +5
Reinforcement learning (RL) has emerged as an effective method for enhancing the reasoning capabilities of large language models (LLMs). However, existing methods suffer from insuf…
Procedural Memory Distillation: Online Reflection for Self-Improving Language Models
Ye Liu, Srijan Bansal, Bo Pang +6
Reinforcement learning with verifiable rewards (RLVR), along with recent selfdistillation variants such as SDPO, evaluates each rollout against a verifier and updates the policy fr…
SSR: Socratic Self-Refine for Large Language Model Reasoning
Haizhou Shi, Ye Liu, Bo Pang +6
Large Language Models (LLMs) have demonstrated remarkable reasoning abilities, yet existing test-time frameworks often rely on coarse self-verification and self-correction, limitin…
Self-Abstraction from Grounded Experience for Plan-Guided Policy Refinement
Hiroaki Hayashi, Bo Pang, Wenting Zhao +6
Large language model (LLM) based agents are increasingly used to tackle software engineering tasks that require multi-step reasoning and code modification, demonstrating promising…
Reasoning Curriculum: Bootstrapping Broad LLM Reasoning from Math
Bo Pang, Deqian Kong, Silvio Savarese +2
Reinforcement learning (RL) can elicit strong reasoning in large language models (LLMs), yet most open efforts focus on math and code. We propose Reasoning Curriculum, a simple two…
CoDA: Coding LM via Diffusion Adaptation
Haolin Chen, Shiyu Wang, Can Qin +12
Diffusion language models promise bidirectional context and infilling capabilities that autoregressive coders lack, yet practical systems remain heavyweight. We introduce CoDA, a 1…