Zhiyuan Yan

We propose an explainable and extendable framework to enhance deepfake detection via multimodal large-language models.

This paper propose a large-scale, high-quality editing dataset, accompanied by a comprehensive benchmark, an advanced editing model, and an effective edit evaluator.

ChemCoTBench bridges complex chemical reasoning with arithmetic-inspired step-by-step workflows, enabling LLMs to systematically tackle real-world tasks like molecular optimization and reaction prediction.

A Reasoning Benchmark on Visual Perception