List your company

Commonsense Reasoning on SOCIALIQA (dev)

73.8AccuracyCKT
39.417648.343857.2766.1962Jun 4, 2023Nov 22, 2023May 12, 2024Oct 31, 2024Apr 21, 2025Oct 10, 2025Mar 31, 2026

Evaluation results

MethodLinks
CKTModel Scale=large2023.0673.8
CALMModel Scale=large2023.0672.7
T5Model Scale=large2023.0671.3
FusionGated-FIRNetDiscovery Source=AI Discovered2026.0342.78
PathGate-FusionNetDiscovery Source=AI Discovered2026.0342.37
Hier-GateNetDiscovery Source=AI Discovered2026.0342.07
AdaMulti-PathGateNetDiscovery Source=AI Discovered2026.0342.07
Mamba2Discovery Source=Human Discovered2026.0341.81
Content-SharpRouterDiscovery Source=AI Discovered2026.0341.81
Gated DeltaNetDiscovery Source=Human Discovered2026.0341.71
DeltaNetDiscovery Source=Human Discovered2026.0340.74