This paper introduces GPT-Fathom, an open-source evaluation suite for large language models (LLMs). It evaluates 10+ LLMs on 20+ benchmarks, providing insights into the evolution from GPT-3 to GPT-4 and improving transparency.

https://arxiv.org/abs//2309.16583

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