notesum.ai

Published at November 21

UnifiedCrawl: Aggregated Common Crawl for Affordable Adaptation of LLMs on Low-Resource Languages

cs.CL

cs.AI

Released Date: November 21, 2024

Authors: Bethel Melesse Tessema¹, Akhil Kedia², Tae-Sun Chung¹

Aff.: ¹Ajou University, Suwon, South Korea; ²Independent Researcher, Seoul, South Korea

Arxiv: http://arxiv.org/abs/2411.14343v1

Models	F1	EM
XGLM- $564\mathrm{M}$	0	0
XGLM- $564\mathrm{M}$ (ours)	0	0
XGLM- $4.5\mathrm{B}$	8.0	1.3
XGLM- $4.5\mathrm{B}$ (ours)	9.9	2.3