565 citations · 691 across the 38 of their papers we have counts for
8 papers · 1 filter
Reuse your FLOPs: Scaling RL on Hard Problems by Conditioning on Very Off-Policy Prefixes
Amrith Setlur, Zijian Wang, Andrew Cohen +2
Typical reinforcement learning (RL) methods for LLM reasoning waste compute on hard problems, where correct on-policy traces are rare, policy gradients vanish, and learning stalls.…
PaperGuide: Making Small Language-Model Paper-Reading Agents More Efficient
Zijian Wang, Tiancheng Huang, Hanqi Li +3
The accelerating growth of the scientific literature makes it increasingly difficult for researchers to track new advances through manual reading alone. Recent progress in large la…
Empowering Multi-Turn Tool-Integrated Agentic Reasoning with Group Turn Policy Optimization
Yifeng Ding, Hung Le, Songyang Han +5
Training Large Language Models (LLMs) for multi-turn Tool-Integrated Reasoning (TIR) - where models iteratively reason, generate code, and verify through execution - remains challe…
Controllable Flow Matching for Online Reinforcement Learning
Bin Wang, Boxiang Tao, Haifeng Jing +2
Model-based reinforcement learning (MBRL) typically relies on modeling environment dynamics for data efficiency. However, due to the accumulation of model errors over long-horizon…
Learn A Flexible Exploration Model for Parameterized Action Markov Decision Processes
Zijian Wang, Bin Wang, Mingwen Shao +2
Hybrid action models are widely considered an effective approach to reinforcement learning (RL) modeling. The current mainstream method is to train agents under Parameterized Actio…
Learning Code Preference via Synthetic Evolution
Jiawei Liu, Thanh Nguyen, Mingyue Shang +5
Large Language Models (LLMs) have recently demonstrated remarkable coding capabilities. However, assessing code generation based on well-formed properties and aligning it with deve…
Planning-Aware Code Infilling via Horizon-Length Prediction
Yifeng Ding, Hantian Ding, Shiqi Wang +3
Fill-in-the-Middle (FIM), or infilling, has become integral to code language models, enabling generation of missing code given both left and right contexts. However, the current FI…
CrossCodeEval: A Diverse and Multilingual Benchmark for Cross-File Code Completion
Yangruibo Ding, Zijian Wang, Wasi Uddin Ahmad +8
Code completion models have made significant progress in recent years, yet current popular evaluation datasets, such as HumanEval and MBPP, predominantly focus on code completion t…