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

Deception Detection on DeceptArena (test)

0.927False Assertion Score

Hybrid

Updated 2mo ago

Evaluation Results

Method	Links
Hybrid 2026.03		0.927	0.904	0.911	0.916	0.947	0.919	0.931	0.942	0.912	0.921	0.908	0.901	0.934
AP-H 2026.03		0.901	0.869	0.877	0.894	0.928	0.879	0.891	0.921	0.867	0.878	0.858	0.847	0.901
CA-P 2026.03		0.884	0.841	0.856	0.889	0.912	0.851	0.868	0.897	0.821	0.834	0.809	0.796	0.869
BB 2026.03		0.861	0.798	0.812	0.874	0.901	0.822	0.843	0.889	0.756	0.779	0.741	0.718	0.837