Zhaoran Wang

Beyond Markovian: Reflective Exploration via Bayes-Adaptive RL for LLM Reasoning

Learning to Reason as Action Abstractions with Scalable Mid-Training RL

Local Linear Attention: An Optimal Interpolation of Linear and Softmax Attention For Test-Time Regression

Are Transformers Able to Reason by Connecting Separated Knowledge in Training Data?

Reward-Augmented Data Enhances Direct Preference Alignment of LLMs

BRiTE: Bootstrapping Reinforced Thinking Process to Enhance Language Model Reasoning

Reward-Augmented Data Enhances Direct Preference Alignment of LLMs

The Sample Complexity of Online Strategic Decision Making with Information Asymmetry and Knowledge Transportability

Progressive LLM Alignments Using Two-Player Games

Provably Efficient and Practical Self-Play for Better LLM Alignment

How Can LLM Guide RL? A Value-Based Approach

An Instrumental Value for Data Production and its Application to Data Pricing

Human-Instruction-Free LLM Self-Alignment with Limited Samples

Hindsight Planner: A Closed-loop few-shot planner for Embodied Instruction Following

Self-Exploring Language Models: Active Preference Elicitation for Online Alignment