List your company

Multiple-Choice Question Answering on MMLU Social Sciences (test)

79.12AccuracyKAPPA
70.144872.474974.80577.1351Sep 28, 2025

Evaluation results

MethodLinks
KAPPAModel=Qwen 2.5 7B, Intervention Layers=62025.0979.1295.470.759
Knowledge Probe (KP)Model=Qwen 2.5 7B2025.0979.021000
KAPPAModel=Qwen 2.5 7B, Intervention Layers=32025.0978.9596.780.739
KAPPAModel=Qwen 2.5 7B, Intervention Layers=12025.0978.9197.120.734
BaseModel=Qwen 2.5 7B2025.0978.8595.930.718
CAAModel=Qwen 2.5 7B2025.0978.8195.940.718
DoLAModel=Qwen 2.5 7B2025.0978.5294.810.725
BaseModel=Llama 3.1 8B2025.0974.0490.550.471
CAAModel=Llama 3.1 8B2025.0973.9390.620.466
KAPPAModel=Llama 3.1 8B, Intervention Layers=12025.0973.7996.080.617
KAPPAModel=Llama 3.1 8B, Intervention Layers=32025.0973.2494.540.607
DoLAModel=Llama 3.1 8B2025.0972.25840.571
KAPPAModel=Llama 3.1 8B, Intervention Layers=62025.0970.4985.980.778