List your company

Confidence Calibration in Retrieval-Augmented Generation on NQ k=5 OOD (test)

0.248ECENAACL
0.239760.295380.3510.40662Jan 16, 2026

Evaluation results

MethodLinks
NAACLBackbone=Qwen2.5-7B-Instruct, Number of retrieved passages (k)=52026.010.2480.75
NAACLBackbone=Llama-3.1-8B-Instruct, Number of retrieved passages (k)=52026.010.2650.674
Label-only SFTBackbone=Llama-3.1-8B-Instruct, Number of retrieved passages (k)=52026.010.2730.653
NAACLBackbone=DeepSeek-R1-Distill-Llama-8B, Number of retrieved passages (k)=52026.010.2760.628
Noise-awareBackbone=Llama-3.1-8B-Instruct, Number of retrieved passages (k)=52026.010.2890.667
Noise-awareBackbone=DeepSeek-R1-Distill-Llama-8B, Number of retrieved passages (k)=52026.010.290.605
Noise-awareBackbone=Qwen2.5-7B-Instruct, Number of retrieved passages (k)=52026.010.3040.641
CoTBackbone=Qwen2.5-7B-Instruct, Number of retrieved passages (k)=52026.010.3130.696
VanillaBackbone=Qwen2.5-7B-Instruct, Number of retrieved passages (k)=52026.010.3220.706
EnsembleBackbone=Qwen2.5-7B-Instruct, Number of retrieved passages (k)=52026.010.3250.694
Label-only SFTBackbone=DeepSeek-R1-Distill-Llama-8B, Number of retrieved passages (k)=52026.010.3290.604
EnsembleBackbone=Llama-3.1-8B-Instruct, Number of retrieved passages (k)=52026.010.3340.693
Label-only SFTBackbone=Qwen2.5-7B-Instruct, Number of retrieved passages (k)=52026.010.3350.667
NAACLBackbone=DeepSeek-R1-Distill-Qwen-7B, Number of retrieved passages (k)=52026.010.3350.64
EnsembleBackbone=DeepSeek-R1-Distill-Llama-8B, Number of retrieved passages (k)=52026.010.3510.59
CoTBackbone=Llama-3.1-8B-Instruct, Number of retrieved passages (k)=52026.010.3520.67
VanillaBackbone=Llama-3.1-8B-Instruct, Number of retrieved passages (k)=52026.010.3710.645
CoTBackbone=DeepSeek-R1-Distill-Llama-8B, Number of retrieved passages (k)=52026.010.3730.621
VanillaBackbone=DeepSeek-R1-Distill-Llama-8B, Number of retrieved passages (k)=52026.010.3760.625
Noise-awareBackbone=DeepSeek-R1-Distill-Qwen-7B, Number of retrieved passages (k)=52026.010.3760.477
EnsembleBackbone=DeepSeek-R1-Distill-Qwen-7B, Number of retrieved passages (k)=52026.010.3980.597
Label-only SFTBackbone=DeepSeek-R1-Distill-Qwen-7B, Number of retrieved passages (k)=52026.010.4020.686
VanillaBackbone=DeepSeek-R1-Distill-Qwen-7B, Number of retrieved passages (k)=52026.010.4350.608
CoTBackbone=DeepSeek-R1-Distill-Qwen-7B, Number of retrieved passages (k)=52026.010.4540.627