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

Steering on Principle-based evaluation dataset

8.68G Score

Claude 4.5-Haiku

7.85848.07178.2858.4983Dec 2, 2025
Updated 3mo ago

Evaluation Results

MethodLinks
2025.12
8.688.768.138.528.678.53
2025.12
8.68.768.098.528.618.6
2025.12
8.538.517.788.338.518.3
2025.12
8.498.597.858.518.618.38
2025.12
8.478.457.888.368.488.35
2025.12
8.48.427.848.318.318.18
2025.12
8.368.457.888.338.358.3
2025.12
8.348.327.958.38.288.43
2025.12
8.38.357.868.238.157.95
2025.12
8.138.257.578.058.227.85
2025.12
88.247.467.898.097.8
2025.12
7.898.147.37.88.047.68