List your company

Question Answering on ConFiQA-MC (Held-out)

52.8AccuracyLODO SABER
23.05630.77838.546.222May 11, 2026

Evaluation results

MethodLinks
LODO SABERBackbone=Qwen-2.5-7B, Setting=Zero-shot2026.0552.888.915.918.5
LODO SABERBackbone=Qwen-2.5-3B, Setting=Zero-shot2026.0552.887.728.728.8
LODO SABERBackbone=Llama-3.1-8B, Setting=Zero-shot2026.0550.782.719.620.7
LODO SABERBackbone=Llama-3.2-3B, Setting=Zero-shot2026.0547.4795.815.9
Ctx-EvalBackbone=Llama-3.2-3B, Setting=Zero-shot2026.0541.663——
Ctx-EvalBackbone=Qwen-2.5-7B, Setting=Zero-shot2026.0536.870.4——
Exp-SCRBackbone=Llama-3.1-8B, Setting=Zero-shot2026.0531.262——
Ctx-EvalBackbone=Qwen-2.5-3B, Setting=Zero-shot2026.0524.258.8——