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

Scalable Oversight on SWE-bench Lite

6.27Misalignment Rate

CCO

5.73969.319812.916.4802May 27, 2026
Updated 6d ago

Evaluation Results

MethodLinks
2026.05
6.271.2778.766.47
2026.05
10.870.8763.2846.67
2026.05
15.670.6746.2934.27
2026.05
19.530.4732.7621.13