Tao Fang

Research paper on NLP model interpretability withdrawn from arXiv

A research paper titled "CLIF: Concept-Level Influence Functions for Transparent Bottleneck Models" was submitted to arXiv, proposing a novel method to enhance the interpretability of NLP models. The approach uses influ…