Llama 2
Meta
partial detail 15 method fields published
Evaluation
Reward model evaluation (Meta Helpfulness and Safety test sets)
Performance of the final helpfulness and safety reward models on a diverse set of human preference benchmarks (Meta Helpfulness and Safety test sets), per-preference-rating granular accuracy, and scaling trends with data and model size
How it was run
Open a release to see the setup, scoring, and other details that the lab published. Only fields the lab actually disclosed are shown.