List your company

Language Understanding on LLM Evaluation Harness

22.9ARC-Challenge AccBased
17.90819.20420.521.796Feb 16, 2024Feb 27, 2024Mar 9, 2024Mar 20, 2024Mar 31, 2024Apr 11, 2024Apr 23, 2024

Evaluation results

MethodLinks
BasedArchitecture=Based, Shot=1-shot2024.0222.9—40.4————50.462.1—30.8———41.3
ReBasedArchitecture=ReBased, Shot=1-shot2024.0221.5—41————54.662.8—30.7———42.1
MH-MoEExperts Number (N)=322024.0421.426.850.644.853.418.831.653.869.356.63542.124.839.540.6
MH-MoEExperts Number (N)=82024.0419.625.250.242.25318.230.351.168.759.633.240.324.740.939.8
X-MoEExperts Number (N)=322024.0419.424.850.442.552.717.83051.368.852.833.440.124.339.139.1
X-MoEExperts Number (N)=82024.041924.748.34252.717.429.850.367.958.431.435.724.340.238.7
Dense2024.0418.123.344.939.751.517.12948.266.65529.734.124.139.337.2