List your company
ResearchFormerly WizWand

LLM steering evaluation on ToneBank

0.37LLM Judge ScoreK-Steering
0.0580.1390.220.301May 30, 2025
Updated 6mo ago

Evaluation results

MethodLinks
K-SteeringModel=Mistral-7B, K=12025.050.37—
CAAModel=Mistral-7B, K=12025.050.36—
CAAModel=Llama-3.2-3B, K=32025.050.33—
CAAModel=Mistral-7B, K=32025.050.3—
K-SteeringModel=Llama-3.2-3B, K=32025.050.29—
K-SteeringModel=Olmo-2-7B, K=32025.050.28—
K-SteeringModel=Llama-3.2-3B, K=12025.050.28—
CAAModel=Mistral-7B, K=22025.050.27—
K-SteeringModel=Mistral-7B, K=32025.050.27—
K-SteeringModel=Olmo-2-7B, K=22025.050.26—
CAAModel=Olmo-2-7B, K=32025.050.26—
K-SteeringModel=Mistral-7B, K=22025.050.25—
K-SteeringModel=Llama-3.2-3B, K=22025.050.24—
CAAModel=Llama-3.2-3B, K=22025.050.23—
K-SteeringModel=Olmo-2-7B, K=12025.050.22—
CAAModel=Olmo-2-7B, K=12025.050.16—
CAAModel=Olmo-2-7B, K=22025.050.14—
CAAModel=Llama-3.2-3B, K=12025.050.07—