← all datasets

enwiki-8

Emerging
1papers using it
2023first seen

'enwiki-8' is a dataset that contains a subset of English Wikipedia text, used to evaluate the performance and efficiency of language models, particularly in the context of resource-efficient architectures like sparse Mixtures of Experts.

Papers using enwiki-8 (1)

enwiki-8 dataset β€” papers, benchmarks & downloads Β· AI for Code