Perplexed by Perplexity: Perplexity-Based Data Pruning With Small Reference Models - ML Notes

Skip to article frontmatter Skip to article content

Overview¶

Introduces a method for pruning datasets based on perplexity measures.

Links¶

Paper
Tags: Data pruning, Perplexity

Notes¶

This is a bunch of stuff that should be put in a note somewhere

References¶

Ankner, Z., Blakeney, C., Sreenivasan, K., Marion, M., Leavitt, M. L., & Paul, M. (2024). Perplexed by Perplexity: Perplexity-Based Data Pruning With Small Reference Models. arXiv. 10.48550/ARXIV.2405.20541