List your company

Language Understanding on MMLU (ΔMMLU)

0.92ΔMMLUFLOATDOOR
-0.02640.21930.4650.7107Jun 17, 2026

Evaluation results

MethodLinks
FLOATDOORScenario=M2, Backbone=Qwen3-4B2026.060.92
FLOATDOORScenario=M1, Backbone=Qwen3-4B2026.060.91
FLOATDOORScenario=S3, Backbone=Qwen3-4B2026.060.24
FLOATDOORScenario=S1, Backbone=Qwen3-4B2026.060.12
FLOATDOORScenario=S2, Backbone=Qwen3-4B2026.060.01