← all datasets

FinerWeb-10BT

Emerging
1papers using it
2025first seen

FinerWeb-10BT is a quality-annotated dataset derived from a 10 billion-token subset of FineWeb, used to evaluate the effectiveness of LLM-based line-level filtering methods for improving data quality in training Large Language Models.

Papers using FinerWeb-10BT (1)

FinerWeb-10BT dataset β€” papers, benchmarks & downloads Β· AI for Code