Disambiguation and completeness on AmbigQA
0.113Personalization BiasIdentity-Robust Generation
Evaluation results
| Method | Links | |
|---|---|---|
| Identity-Robust GenerationModel=Llama3.3-70B2026.01 | 0.113 | |
| Identity-Robust GenerationModel=Qwen32026.01 | 0.131 | |
| Identity-Robust GenerationModel=gpt-oss-20B2026.01 | 0.171 | |
| Prompt SteeringModel=Llama3.3-70B2026.01 | 0.337 | |
| Vanilla GenerationModel=Llama3.3-70B2026.01 | 0.402 | |
| Prompt SteeringModel=gpt-oss-20B2026.01 | 0.467 | |
| Vanilla GenerationModel=Qwen32026.01 | 0.47 | |
| Prompt SteeringModel=Qwen32026.01 | 0.557 | |
| Vanilla GenerationModel=gpt-oss-20B2026.01 | 0.645 |