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

Instruction Following on IFEval (Precision, Intent, and Utility Scores)

84.1Precision (Short)

SPLoRA

56.22863.46470.777.936May 28, 2026
Updated 2d ago

Evaluation Results

MethodLinks
2026.05
84.188.4985.0389.4586.772.35
2026.05
81.8987.4184.4789.3385.781.36
2026.05
81.5286.9383.1888.37850.58
2026.05
80.9686.6983.7388.8585.060.64
2026.05
80.9686.5782.2687.7784.39-0.03
2026.05
80.4186.5782.4488.2584.420
2026.05
79.385.1380.9686.4582.960
2026.05
77.6383.8179.385.6181.59-1.37
2026.05
77.2684.0579.6786.0981.77-1.19
2026.05
76.7183.0979.6785.3781.21-1.75
2026.05
76.5283.5778.5684.8980.89-2.07
2026.05
74.1281.5377.0884.0579.2-3.76
2026.05
69.6978.974.6882.7376.5-7.92
2026.05
66.5475.5472.0980.173.57-9.39
2026.05
60.2671.2262.1172.366.47-16.49
2026.05
57.367.9959.3369.6663.57-20.85