Cross-lingual Cultural Consistency on BLEnD Indo-European
0.021Max SigmaVanilla
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| VanillaModel=Qwen-2.5-3B2026.04 | 0.021 | 0.244 | — | — | — | |
| Few-shotModel=Qwen-2.5-3B2026.04 | 0.021 | 0.28 | — | — | — | |
| SteeringModel=Qwen-2.5-3B2026.04 | 0.021 | — | 0.239 | 0.248 | 0.254 | |
| PersonaModel=Qwen-2.5-3B2026.04 | 0.021 | — | 0.24 | 0.268 | 0.303 | |
| C-3POModel=Qwen-2.5-3B2026.04 | 0.021 | 0.347 | — | — | — | |
| VanillaModel=Gemma-2-27B2026.04 | 0.022 | 0.627 | — | — | — | |
| Few-shotModel=Gemma-2-27B2026.04 | 0.022 | 0.596 | — | — | — | |
| SteeringModel=Gemma-2-27B2026.04 | 0.022 | — | 0.62 | 0.624 | 0.627 | |
| PersonaModel=Gemma-2-27B2026.04 | 0.022 | — | 0.625 | 0.652 | 0.679 | |
| C-3POModel=Gemma-2-27B2026.04 | 0.022 | 0.674 | — | — | — | |
| VanillaModel=Llama-3.1-8B2026.04 | 0.023 | 0.426 | — | — | — | |
| Few-shotModel=Llama-3.1-8B2026.04 | 0.023 | 0.432 | — | — | — | |
| SteeringModel=Llama-3.1-8B2026.04 | 0.023 | — | 0.419 | 0.431 | 0.441 | |
| PersonaModel=Llama-3.1-8B2026.04 | 0.023 | — | 0.408 | 0.447 | 0.496 | |
| C-3POModel=Llama-3.1-8B2026.04 | 0.023 | 0.494 | — | — | — |