Moral Permissibility On Big Bench

Accuracy

Results

Performance results of various models on this benchmark

Comparison Table

Model Name	Accuracy
scaling-language-models-methods-analysis-1	55.1
training-compute-optimal-large-language	57.3