Yuxin Wen

FictionalQA: A Dataset for Studying Memorization and Knowledge Acquisition

RL Is a Hammer and LLMs Are Nails: A Simple Reinforcement Learning Recipe for Strong Prompt Injection

Scalable Pretraining of Retrieval Models

Quantifying Cross-Modality Memorization in Vision-Language Models

Efficient Vision-Language Models by Summarizing Visual Tokens into Compact Registers

GenQA: An Instruction Dataset of LLM Generated Questions and Answers

Extracting Parameter Counts from Fine-Tuning APIs