Opened about a year ago
by @yuro-1981
BLiMP offers 67 sub-datasets, each with 1,000 minimal pairs, making it incredibly thorough for linguistic evaluation.
0/10000
Sign up or Log in to comment