Question Answering on ConflictQA (Held-out)
46.5AccuracyLODO SABER
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| LODO SABERBackbone=Llama-3.1-8B, Setting=Zero-shot2026.05 | 46.5 | 65.6 | 5.6 | 2.5 | |
| LODO SABERBackbone=Qwen-2.5-7B, Setting=Zero-shot2026.05 | 46 | 68.3 | 7.8 | 0.7 | |
| LODO SABERBackbone=Llama-3.2-3B, Setting=Zero-shot2026.05 | 43 | 67.9 | 5.8 | -2.8 | |
| LODO SABERBackbone=Qwen-2.5-3B, Setting=Zero-shot2026.05 | 41 | 73.1 | 10.4 | 11.7 | |
| Exp-SCRBackbone=Llama-3.1-8B, Setting=Zero-shot2026.05 | 40.9 | 63.1 | — | — | |
| Int-EvalBackbone=Qwen-2.5-7B, Setting=Zero-shot2026.05 | 38.2 | 67.6 | — | — | |
| Int-EvalBackbone=Llama-3.2-3B, Setting=Zero-shot2026.05 | 37.2 | 70.7 | — | — | |
| Int-EvalBackbone=Qwen-2.5-3B, Setting=Zero-shot2026.05 | 30.6 | 61.4 | — | — |