Xifeng Yan

Learning to Lie: Adversarial Attacks on Human-AI Teams and LLMs

SOPBench: Evaluating Language Agents at Following Standard Operating Procedures and Constraints

Investigating Language Models for Supporting Complex Group Decisions

Direct Multi-token Decoding

Adaptive Layer-skipping in Pre-trained LLMs

Open-Qwen2VL: Compute-Efficient Pre-Training of Fully-Open Multimodal LLMs on Academic Resources

MMSci: A Dataset for Graduate-Level Multi-Discipline Multimodal Scientific Understanding

Can Editing LLMs Inject Harm?