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

Uncertainty Quantification on τ2-bench Retail

0.899AUROC

Sonnet-4.6

Updated 1mo ago

Evaluation Results

Method	Links
Sonnet-4.6 2026.06		0.899
Sonnet-4.6 2026.06		0.864
Sonnet-4.6 2026.06		0.852
Progress Advantage 2026.06		0.69
WildReward-8B 2026.06		0.689
Progress Advantage 2026.06		0.678
ThinkPRM-7B 2026.06		0.67
Progress Advantage 2026.06		0.664
Sonnet-4.6 2026.06		0.656
Progress Advantage 2026.06		0.65
WildReward-8B 2026.06		0.643
ThinkPRM-14B 2026.06		0.637
ThinkPRM-14B 2026.06		0.61
Self-Certainty 2026.06		0.608
DeepConf Tail 2026.06		0.608
WildReward-8B 2026.06		0.584
DeepConf B10 2026.06		0.582
ThinkPRM-14B 2026.06		0.573
ThinkPRM-7B 2026.06		0.551
ThinkPRM-14B 2026.06		0.544
ThinkPRM-7B 2026.06		0.543
DeepConf B10 2026.06		0.496
ThinkPRM-7B 2026.06		0.469
WildReward-8B 2026.06		0.468
DeepConf B10 2026.06		0.416
Self-Certainty 2026.06		0.397
Self-Certainty 2026.06		0.392
DeepConf Tail 2026.06		0.382
DeepConf Tail 2026.06		0.38
Self-Certainty 2026.06		0.366
DeepConf Tail 2026.06		0.344
DeepConf B10 2026.06		0.288