π The First Self-Improving agents with RL / Prompting Optimization
We include the following datasets linked to our best understanding of their original source.
GSM8K
MMLU
Mini Crosswords
GAIA
HumanEval