Hongning Wang

Trust-Region Adaptive Policy Optimization

BARREL: Boundary-Aware Reasoning for Factual and Reliable LRMs

AgentRL: Scaling Agentic Reinforcement Learning with a Multi-Turn, Multi-Task Framework

Be Careful When Fine-tuning On Open-Source LLMs: Your Fine-tuning Data Could Be Secretly Stolen!

Think Socially via Cognitive Reasoning

From Theft to Bomb-Making: The Ripple Effect of Unlearning in Defending Against Jailbreak Attacks

RLAR: An Agentic Reward System for Multi-task Reinforcement Learning on Large Language Models

Beyond Nash Equilibrium: Bounded Rationality of LLMs and humans in Strategic Decision-making

Data Selection via Optimal Control for Language Models

MAPS: Advancing Multi-Modal Reasoning in Expert-Level Physical Science

SPaR: Self-Play with Tree-Search Refinement to Improve Instruction-Following in Large Language Models

RecFlow: An Industrial Full Flow Recommendation Dataset

Does RLHF Scale? Exploring the Effects of Data, Model, and Method

RedHat: Towards Reducing Hallucination in Essay Critiques with Large Language Models

CodePlan: Unlocking Reasoning Potential in Large Language Models by Scaling Code-form Planning

Cost-Efficient Multi-Fidelity Alignment for LLMs