Scientific Reasoning on SciEval

73.93Score

GPT-4

Updated 3d ago

Evaluation Results

Method	Links
GPT-4 2024.01		73.93
Llama3-8B-Instruct 2024.01		71.38
Llama3-8B-Instruct 2024.01		71.38
SciGLM 2024.01		69.84
ChatGLM3-32B-Base 2024.01		67.38
GPT-3.5-turbo 2024.01		66.97
Llama3-8B-Instruct + SciInstruct 2024.01		66.47
Mistral-7B: MetaMATH + SciInstruct 2024.01		64.16
Mistral-7B: MetaMATH 2024.01		63.61
Mistral-7B: MetaMATH 2024.01		63.61
Claude-v1.3 2024.01		63.45
SciGLM 2024.01		62.09
ChatGLM3-6B-Base 2024.01		61.69
ChatGLM3-6B 2024.01		56.56
Vicuna-13B 2024.01		53.93
ChatGLM2-6B 2024.01		53.02
Galactica-6.7B 2024.01		50.87
ChatGLM2-6B-Base 2024.01		50.38
LLaMA-2-13B 2024.01		36.96
LLaMA-2-7B 2024.01		28.37