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

Role-Playing on RoleBench (test)

88.82LLM-as-a-Judge Score

Llama-3.1-8B-Instruct

-2.512821.198644.9168.6214Jun 12, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
2026.06
88.82
2026.06
84.33
2026.06
84.26
2026.06
83.35
2026.06
82.62
2026.06
81.2
2026.06
80.01
2026.06
70.37
2026.06
67.86
2026.06
52.82
2026.06
48.43
2026.06
47.98
2026.06
39.65
2026.06
33.56
2026.06
33.41
2026.06
33.25
2026.06
30.26
2026.06
30
2026.06
27.75
2026.06
27.3
2026.06
25.72
2026.06
25.66
2026.06
25.61
2026.06
24.63
2026.06
23.95
2026.06
20.53
2026.06
17.58
2026.06
16.31
2026.06
14.29
2026.06
13.99
2026.06
11.4
2026.06
10.95
2026.06
9.5
2026.06
6.76
2026.06
5.91
2026.06
3.76
2026.06
1.78
2026.06
1.41
2026.06
1.35
2026.06
1.21
2026.06
1
2026.06
1