test_caa_steering / activation_steering_metadata.json
shreyansjain's picture
Add steering model artifacts
5a32a9b verified
{
"model_name": "unsloth/Llama-3.2-1B-Instruct",
"unique_labels": [
"Analogy Construction",
"Appeal to Precedent",
"Burden of Proof Shift",
"Circular Anticipation",
"Concession and Pivot",
"Empirical Grounding",
"Moral Framing",
"Reductio ad Absurdum",
"Refutation by Distinction",
"Straw Man Reframing"
],
"is_fitted": true,
"class_name": "CAASteering"
}