🌍 MultiBLiMP Leaderboard
Accuracy (%) on MultiBLiMP per language. Mean is macro-average across all languages.
Show language column
0 200
Llama-3.1-70B-Instruct | goldfish-models | 70.6 | 86.4 | 42.5 | 82.4 | 96.4 | 96.4 | 42.9 | 95.7 | 69.1 | 97.2 | 90.5 | 85.3 | 65.6 | 97.3 | 77.7 | 98.7 | 98.3 | 98.1 | 69.5 | 94.5 | 100 | 99.2 | 40.9 | 99.4 | 98.6 | 96.3 | 97.8 | 89.2 | 96.2 | 97.6 | 99.5 | 96.6 | 78.8 | 98.5 | 92.9 | 94.8 | 68.1 | 91.7 | 85.7 | 93.9 | 99.2 | 92.6 | 99.4 | 58 | 82.8 | 99.2 | 98.7 | 97.7 | 97.1 | 96.8 | 95.6 | 97.7 | 99.5 | 93.6 | 72.1 | 77.8 | 73.1 | 52.5 | 91.4 | 97.6 | 78.3 | 98.4 | 90.9 | 56.1 | 94.9 | 66.2 | 77.7 | 52.6 | 98.4 | 91.6 | 81.7 | 92.9 | 92.3 | 96.1 | 98.3 | 75.6 | 98.8 | 99.5 | 86.8 | 79.4 | 97.8 | 95.4 | 78.7 | 77.2 | 98.3 | 98.8 | 99.5 | 97.9 | 46.4 | 95.7 | 85.2 | 98.1 | 53.8 | 97.6 | 100 | 67.4 | 33.3 | 84.4 | 81.1 | 81.4 | 90 | 61.8 | 55.6 |