List your company
ResearchFormerly WizWand

Multiple-choice Question Answering on Anthropic Behavioral Concepts CAA-style (Negative Steering)

8AI CoordinationSAS
3.844.9267.08Jun 13, 2026
Updated 4mo ago

Evaluation results

MethodLinks
SASSteering Method=SAS, Steering Strength=alpha < 0 (minimum over {-2, -1}), System Prompt Condition=no system prompt, Steering Layer=22, Model=GEMMA-2-9B-IT2026.068—2728166744
CAASteering Method=CAA, Steering Strength=alpha < 0 (minimum over {-2, -1}), System Prompt Condition=no system prompt, Steering Layer=22, Model=GEMMA-2-9B-IT2026.066—1924173535
No steeringSteering Method=None, Model=GEMMA-2-9B-IT2026.066—4429246753
HiDRASteering Method=HiDRA, Steering Strength=alpha < 0 (minimum over {-2, -1}), System Prompt Condition=no system prompt, Steering Layer=22, Model=GEMMA-2-9B-IT2026.064—9872927