Question Answering on BBQ Nationality
82.1AccuracyCatRAG
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| CatRAGBackbone=Llama-32026.03 | 82.1 | 0.02 | |
| CatRAGBackbone=GPT-OSS2026.03 | 81.7 | 0.03 | |
| CatRAGBackbone=Gemma-32026.03 | 81.2 | 0.02 | |
| CausalBackbone=GPT-OSS2026.03 | 80.7 | 0.04 | |
| CausalBackbone=Gemma-32026.03 | 78.7 | 0.1 | |
| CausalBackbone=Llama-32026.03 | 78.5 | 0.1 | |
| SPBackbone=GPT-OSS2026.03 | 74 | 0.14 | |
| SPBackbone=Gemma-32026.03 | 71.2 | 0.16 | |
| SPBackbone=Llama-32026.03 | 69.8 | 0.18 | |
| CEBackbone=GPT-OSS2026.03 | 69.6 | 0.22 | |
| CEBackbone=Gemma-32026.03 | 66.8 | 0.25 | |
| CEBackbone=Llama-32026.03 | 65.3 | 0.27 | |
| SelfBackbone=GPT-OSS2026.03 | 64.5 | 0.34 | |
| SelfBackbone=Gemma-32026.03 | 61.5 | 0.37 | |
| SelfBackbone=Llama-32026.03 | 60.2 | 0.39 | |
| BaseBackbone=GPT-OSS2026.03 | 54.5 | 0.54 | |
| BaseBackbone=Gemma-32026.03 | 51.3 | 0.59 | |
| BaseBackbone=Llama-32026.03 | 50.1 | 0.61 |