open_llama
open_llama copied to clipboard
Eval for 300B checkpoint
Excellent work. For the checkpoint at 300B tokens, would you mind sharing the lm-evaluation-harness metrics?