Zeping Yu

icon indicating a website link https://scholar.google.com/citations?hl=zh-CN&user=OdpmpDsAAAAJ

icon company The University of Manchester icon location Manchester, UK icon location mechanistic interpretability of LLM

Results 3 issues of Zeping Yu

SRNN's speed could increase more!

Recently when I read the codes again, I found the "timedistributed" wrapper in keras actually is not computed parallelized, they use a "for loop" when implementing, so the computing over...

how to merge the base model and the fine-tuned model in mergekit?

models: - model: models/vicuna-7b-v1.5-16k parameters: weight: 1.0 - model: models/vicuna_7b_A parameters: weight: 1.0 base_model: models/vicuna-7b-v1.5-16k merge_method: della parameters: normalize: true int8_mask: true density: 0.7 lambda: 1.1 epsilon: 0.2 dtype: float16...

paper recommendation: interpretability analysis of multimodal LLM

title: Understanding Multimodal LLMs: the Mechanistic Interpretability of Llava in Visual Question Answering link: https://arxiv.org/pdf/2411.10950