Gender on Patchscopes (test)
76.09SRBALOR (S)
Evaluation Results
| Method | Links | |
|---|---|---|
| BALOR (S)Backbone=Qwen3-4b, Setting=Zero-shot2026.01 | 76.09 | |
| BALOR (D)Backbone=Llama3.2-1b, Setting=Zero-shot2026.01 | 63.27 | |
| BALOR (S)Backbone=Llama3.2-1b, Setting=Zero-shot2026.01 | 59.94 | |
| BALOR (D)Backbone=Qwen3-4b, Setting=Zero-shot2026.01 | 58.26 | |
| BALOR (S)Backbone=Llama3-8b, Setting=Zero-shot2026.01 | 54.39 | |
| BALOR (D)Backbone=Llama3-8b, Setting=Zero-shot2026.01 | 51.71 | |
| CBBackbone=Llama3-8b, Setting=Zero-shot2026.01 | 50 | |
| IEBackbone=Llama3-8b, Setting=Zero-shot2026.01 | 50 | |
| Fine-tuningBackbone=Qwen3-4b, Setting=Zero-shot2026.01 | 48.58 | |
| DBBackbone=Llama3-8b, Setting=Zero-shot2026.01 | 47.56 | |
| CBBackbone=Llama3.2-1b, Setting=Zero-shot2026.01 | 46.67 | |
| BALOR (S)Backbone=Llama2-7b, Setting=Zero-shot2026.01 | 46.42 | |
| BALOR (D)Backbone=Llama2-7b, Setting=Zero-shot2026.01 | 43.7 | |
| IEBackbone=Llama2-7b, Setting=Zero-shot2026.01 | 41.7 | |
| Fine-tuningBackbone=Llama2-7b, Setting=Zero-shot2026.01 | 41.02 | |
| DBBackbone=Llama2-7b, Setting=Zero-shot2026.01 | 37.92 | |
| CBBackbone=Llama2-7b, Setting=Zero-shot2026.01 | 36.04 | |
| VanillaBackbone=Qwen3-4b, Setting=Zero-shot2026.01 | 34.78 | |
| Fine-tuningBackbone=Llama3-8b, Setting=Zero-shot2026.01 | 32.93 | |
| FairSteerBackbone=Qwen3-4b, Setting=Zero-shot2026.01 | 31.13 | |
| DBBackbone=Llama3.2-1b, Setting=Zero-shot2026.01 | 26.64 | |
| VanillaBackbone=Llama2-7b, Setting=Zero-shot2026.01 | 24.06 | |
| FairSteerBackbone=Llama2-7b, Setting=Zero-shot2026.01 | 22.64 | |
| IEBackbone=Llama3.2-1b, Setting=Zero-shot2026.01 | 9.99 | |
| IEBackbone=Qwen3-4b, Setting=Zero-shot2026.01 | 8.7 | |
| VanillaBackbone=Llama3.2-1b, Setting=Zero-shot2026.01 | 3.33 | |
| Fine-tuningBackbone=Llama3.2-1b, Setting=Zero-shot2026.01 | 3.33 | |
| CBBackbone=Qwen3-4b, Setting=Zero-shot2026.01 | 2.17 | |
| DBBackbone=Qwen3-4b, Setting=Zero-shot2026.01 | 2.17 | |
| VanillaBackbone=Llama3-8b, Setting=Zero-shot2026.01 | 0 | |
| FairSteerBackbone=Llama3.2-1b, Setting=Zero-shot2026.01 | 0 | |
| FairSteerBackbone=Llama3-8b, Setting=Zero-shot2026.01 | 0 |