Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Reading Comprehension on RACE (First-Token Accuracy)

87.3First-Token Accuracy

Zephyr-7B

47.98858.19468.478.606May 21, 2025
Updated 12d ago

Evaluation Results

MethodLinks
2025.05
87.3
2025.05
87
2025.05
86.7
2025.05
86.7
2025.05
86.5
2025.05
86.3
2025.05
86.2
2025.05
84.7
2025.05
84.7
2025.05
83.8
2025.05
83.2
2025.05
83.1
2025.05
82.7
2025.05
79.5
2025.05
78.2
2025.05
78.1
2025.05
77.5
2025.05
73.4
2025.05
73.3
2025.05
72.3
2025.05
70.5
2025.05
65.2
2025.05
50.4
2025.05
49.5