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

Instruction Following on BBH

67.1Accuracy

Baseline Step 0

7.92423.28738.6554.013Feb 5, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
2026.02
67.11
2026.02
57.61
2026.02
57.32.4
2026.02
56.82.6
2026.02
56.43
2026.02
56.23.2
2026.02
563.3
2026.02
55.63.4
2026.02
53.83.9
2026.02
51.94.1
2026.02
51.42
2026.02
50.44.2
2026.02
45.54.4
2026.02
45.13
2026.02
39.44.6
2026.02
34.41
2026.02
33.64.8
2026.02
284
2026.02
281
2026.02
27.65.1
2026.02
27.42.1
2026.02
27.22.3
2026.02
26.92
2026.02
26.82.5
2026.02
26.42.6
2026.02
26.22.6
2026.02
26.22.7
2026.02
26.13
2026.02
25.83.1
2026.02
25.23.1
2026.02
24.63.3
2026.02
23.53.4
2026.02
22.53.6
2026.02
21.83.7
2026.02
21.25.3
2026.02
213
2026.02
19.74
2026.02
16.65
2026.02
14.94
2026.02
10.25