List your company

Physical Reasoning on PROST

29.6AccuracyGPT-NeoX
21.17623.36325.5527.737Apr 14, 2022Oct 24, 2022May 5, 2023Nov 14, 2023May 25, 2024Dec 4, 2024Jun 15, 2025

Evaluation results

MethodLinks
GPT-NeoXModel Size=20B, Zero-shot=true2022.0429.6
MaskProBase Model=DeepSeek-7B, Sparsity=2:42025.0629.57
GPT-3Model Variant=Curie, Zero-shot=true2022.0428.8
SparsegptBase Model=DeepSeek-7B, Sparsity=2:42025.0628.22
DenseBase Model=DeepSeek-7B, Sparsity=Dense2025.0628.19
Pruner-ZBase Model=DeepSeek-7B, Sparsity=2:42025.0627.62
FairSeqNumber of Parameters=6.7B, Zero-Shot=true2022.0427.2
GPT-3Model Variant=Babbage, Zero-shot=true2022.0427
GPT-JModel Size=6B, Zero-shot=true2022.0426.7
GPT-3Model Variant=DaVinci, Zero-shot=true2022.0426.7
FairSeqNumber of Parameters=355M, Zero-Shot=true2022.0425.7
FairSeqNumber of Parameters=1.3B, Zero-Shot=true2022.0425.7
GPT-3Model Variant=Ada, Zero-shot=true2022.0425.4
FairSeqNumber of Parameters=13B, Zero-Shot=true2022.0425.2
MaskProBase Model=LLaMA-2-7B, Sparsity=2:42025.0624.41
SparsegptBase Model=LLaMA-2-7B, Sparsity=2:42025.0624.26
Pruner-ZBase Model=LLaMA-2-7B, Sparsity=2:42025.0624.03
DenseBase Model=LLaMA-2-7B, Sparsity=Dense2025.0623.6
FairSeqNumber of Parameters=2.7B, Zero-Shot=true2022.0423
FairSeqNumber of Parameters=125M, Zero-Shot=true2022.0421.5