The paper introduces a novel pruning method, Wanda, for inducing sparsity in pretrained Large Language Models (LLMs) without requiring retraining or weight update, and evaluates its performance on various language benchmarks.

https://arxiv.org/abs//2306.11695

YouTube: https://www.youtube.com/@ArxivPapers

PODCASTS:

Apple Podcasts: https://podcasts.apple.com/us/podcast/arxiv-papers/id1692476016

Spotify: https://podcasters.spotify.com/pod/show/arxiv-papers