Share your thoughts, 1 month free Claude Pro on usSee more

Question Answering with Abstention on SimpleQA

0.2BAS

MBR

Updated 2mo ago

Evaluation Results

Method	Links
MBR 2026.05		0.2
MBR 2026.05		0.19
MBR 2026.05		0.18
Beam 2026.05		0.17
Beam 2026.05		0.17
Greedy 2026.05		0.16
Greedy 2026.05		0.16
Contrastive 2026.05		0.15
Contrastive 2026.05		0.14
Greedy 2026.05		0.14
Beam 2026.05		0.13
Contrastive 2026.05		0.11
MBR 2026.05		0.07
Greedy 2026.05		0.04
Beam 2026.05		0.03
Contrastive 2026.05		0.03
Task Calibration 2026.05		0.03
Task Calibration 2026.05		0
Task Calibration 2026.05		0
Task Calibration 2026.05		0
Ancestral 2026.05		-0.01
Ancestral 2026.05		-0.1
Ancestral 2026.05		-0.1
Ancestral 2026.05		-0.11