Mingxuan Wang

MemAgent: Reshaping Long-Context LLM with Multi-Conv RL-based Memory Agent

Scaling Linear Attention Capacity with Sparse State Expansion

AetherCode: Evaluating LLMs’ Ability to Win In Premier Programming Competitions

ThinkDial: An Open Recipe for Controlling Reasoning Effort in Large Language Models

Truncated Proximal Policy Optimization

Enigmata: Scaling Logical Reasoning in Large Language Models with Synthetic Verifiable Puzzles

ShortListing Model: A Streamlined Simplex Diffusion for Discrete Variable Generation

DAPO: An Open-Source LLM Reinforcement Learning System at Scale

Let Large Language Models Find the Data to Train Themselves