cx-cmu/repro-rl-data

Dataset

This is the RL training data for repro-rephraser-4B from RePro: Training Language Models to Faithfully Recycle the Web for Pretraining.

0

4 commits

1 linked in READMEs

updated Oct 18, 2025

See the code

README

This is the RL training data for repro-rephraser-4B from RePro: Training Language Models to Faithfully Recycle the Web for Pretraining.

Each entry contains two columns, dataman_score and text.

Code: https://github.com/cxcscmu/RePro