Yoshi Suhara

This article examines biases and inconsistencies in LLM evaluations and suggests improvement methods.

Research shows simple classifiers can spot unique patterns in AI-generated text.

A method to shrink language models without sacrificing effectiveness through pruning and distillation.