PPL
Yuchen Zhang
Papers (8)
Tarsier: Recipes for Training and Evaluating Large Video Description ModelsOmniVL-Guard Pro: A Tool-Augmented Agent for Omnibus Vision-Language ForensicsDynaTrain: Fast Online Parallelism Switching for Elastic LLM TrainingContinuous Discovery of Vulnerabilities in LLM Serving Systems with FuzzingMT-JailBench: A Modular Benchmark for Understanding Multi-Turn Jailbreak AttacksGESR: Graph-Based Edge Semantic Reconstruction for Stealthy Communication Detection with Benign-Only TrainingImplicit Compression Regularization: Concise Reasoning via Internal Shorter Distributions in RL Post-TrainingTeaching Thinking Models to Reason with Tools: A Full-Pipeline Recipe for Tool-Integrated Reasoning