Llama 1
Meta
partial detail 11 method fields published
Evaluation
Common sense reasoning benchmarks (BoolQ, PIQA, SIQA, HellaSwag, WinoGrande, ARC, OpenBookQA, CommonsenseQA)
Zero-shot performance on eight standard common sense reasoning benchmarks (Cloze and Winograd-style tasks, multiple choice QA)
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.