Concept Steering on AxBench (Held-out)
1.113HMeanFLAS
Evaluation Results
| Method | Links | |
|---|---|---|
| FLASBase Model=Gemma-2-9B-IT, Intervention Layer=Layer 20, T=22026.05 | 1.113 | |
| PromptingBase Model=Gemma-2-9B-IT, Intervention Layer=Layer 202026.05 | 1.091 | |
| FLASBase Model=Gemma-2-2B-IT, Intervention Layer=Layer 20, T=22026.05 | 1.015 | |
| HyperSteerBase Model=Gemma-2-9B-IT, Intervention Layer=Layer 202026.05 | 0.934 | |
| PromptingBase Model=Gemma-2-2B-IT, Intervention Layer=Layer 202026.05 | 0.762 | |
| HyperSteerBase Model=Gemma-2-2B-IT, Intervention Layer=Layer 202026.05 | 0.608 |