Bowen Yu

WorldPM: Understanding Scaling Patterns in Human Preference Modeling

Beyond Turn Limits: Training Deep Search Agents with Dynamic Context Window

RefCritic: Training Long Chain-of-Thought Critic Models with Refinement Feedback

Beyond the 80/20 Rule: High-Entropy Minority Tokens Drive Effective Reinforcement Learning for LLM Reasoning

Self-play with Execution Feedback: Improving Instruction-following Capabilities of Large Language Models

MARGE: Improving Math Reasoning with Guided Exploration

Self-Evolving Critique Abilities in Large Language Models

Efficient Long Context Fine-tuning with Chunk Flow

Extend Model Merging from Fine-Tuned to Pre-Trained Large Language Models via Weight Disentanglement

Rethinking Data Selection at Scale: Random Selection is Almost All You Need

Language Models can Self-Lengthen to Generate Long Texts

Aligning Large Language Models via Self-Steering Optimization

A Unified View of Delta Parameter Editing in Post-Trained Large-Scale Models