Yixuan Li

Detecting Distillation Data from Reasoning Models

Perturbations Matter: Sensitivity-Guided Hallucination Detection in LLMs

Combating Data Laundering in LLM Training

Measuring and Mitigating Identity Bias in Multi-Agent Debate via Anonymization

GeoArena: An Open Platform for Benchmarking Large Vision-language Models on WorldWide Image Geolocalization

Cognition-of-Thought Elicits Social-Aligned Reasoning in Large Language Models

OOD Learner via In-Context Learning

Unsupervised Domain Adaptation with an Unobservable Source Subpopulation

Visionary-R1: Mitigating Shortcuts in Visual Reasoning with Reinforcement Learning

Do Vision Language Models Rotate in Mind? Evaluating Spatial Transformation Reasoning

Unlocking the Pre-Trained Model as a Dual-Alignment Calibrator for Post-Trained LLMs

Your Weak LLM is Secretly a Strong Teacher for Alignment

How Contaminated Is Your Benchmark? Measuring Dataset Leakage in Large Language Models with Kernel Divergence

DaWin: Training-free Dynamic Weight Interpolation for Robust Adaptation

Process Reward Model with Q-value Rankings

Steer LLM Latents for Hallucination Detection

Unsolvable Problem Detection: Evaluating Trustworthiness of Large Multimodal Models

Understanding and Mitigating Miscalibration in Prompt Tuning for Vision-Language Models

CONDA: Adaptive Concept Bottleneck for Foundation Models Under Distribution Shifts

Understanding Multimodal LLMs Under Distribution Shifts: An Information-Theoretic Approach

Understanding and Mitigating Miscalibration in Prompt Tuning for Vision-Language Models

On the Generalization of Preference Learning with DPO

CodeLutra: Boosting LLM Code Generation via Preference-Guided Refinement

VLMGuard: Defending VLMs against Malicious Prompts via Unlabeled Data

Understanding Generalization of Preference Optimization Under Noisy Feedback

How Reliable Is Human Feedback For Aligning Large Language Models?