Ming Jin

AccidentBench: Benchmarking Multimodal Understanding and Reasoning in Vehicle Accidents and Beyond

Adversarial Déjà Vu: Jailbreak Dictionary Learning for Stronger Generalization to Unseen Attacks

StyleBench: Evaluating thinking styles in Large Language Models

Don’t Trade Off Safety: Diffusion Regularization for Constrained Offline RL

Reinforcement Learning with Backtracking Feedback

Robust Gymnasium: A Unified Modular Benchmark for Robust Reinforcement Learning

LLMs Can Plan Only If We Tell Them

Probing Hidden Knowledge Holes in Unlearned LLMs

LLMs Can Reason Faster Only If We Let Them

A Black Swan Hypothesis: The Role of Human Irrationality in AI Safety

Just Enough Shifts: Mitigating Over-Refusal in Aligned Language Models with Targeted Representation Fine-Tuning

Data-Centric Human Preference Optimization with Rationales

LLM Spark: Critical Thinking Evaluation of Large Language Models

Actions Speak Louder Than States: Going Beyond Bayesian Inference in In-Context Reinforcement Learning

Rethinking the Uncertainty: A Critical Review and Analysis in the Era of Large Language Models