Xiaojun Quan

SPELL: Self-Play Reinforcement Learning for Evolving Long-Context Language Models

FuseRL: Dense Preference Optimization for Heterogeneous Model Fusion

Smooth Gradients, Stable Learning: Logits Convexity for Reinforcement Learning

Advantage-Guided Distillation for Preference Alignment in Small Language Models

Probabilistic Token Alignment for Large Language Model Fusion

Lookahead Routing for Large Language Models

FuseChat: Knowledge Fusion of Chat Models

Probabilistic Token Alignment for Large Language Model Fusion

Weighted-Reward Preference Optimization for Implicit Model Fusion

Discriminative Policy Optimization for Token-Level Reward Models

Let Large Language Models Find the Data to Train Themselves