Mistral 7B

Benchmark Model Rank Results
answerability-prediction-on-peerqaMistral-IT-v02-7B-32k#1Macro F1: 0.4703
arithmetic-reasoning-on-gsm8kMistral 7B (maj@8)#94Accuracy: 52.2Parameters (Billion): 7
code-generation-on-mbppMistral 7B (3-shot)#57Accuracy: 47.5
common-sense-reasoning-on-arc-challengeMistral 7B (0-shot)#19Accuracy: 55.5
common-sense-reasoning-on-arc-easyMistral 7B (0-shot)#14Accuracy: 80.0
common-sense-reasoning-on-winograndeMistral 7B (0-shot)#25Accuracy: 75.3
math-word-problem-solving-on-mathMistral 7B (maj@4)#97Accuracy: 13.1Parameters (Billions): 7
multi-task-language-understanding-on-mmluMistral 7B (5-shot)#13Average (%): 60.1
question-answering-on-natural-questionsMistral 7B (5-shot)#29EM: 28.8
question-answering-on-peerqaMistral-v02-7B-32k#2Prometheus-2 Answer Correctness: 3.4245Rouge-L: 0.1922
question-answering-on-piqaMistral 7B (0-shot)#15Accuracy: 83.0
question-answering-on-triviaqaMistral 7B (5-shot)#18EM: 69.9
sentence-completion-on-hellaswagMistral 7B (0-shot)#32Accuracy: 81.3
zero-shot-video-question-answer-on-intentqaMistral (7B)#11Accuracy: 50.4
zero-shot-video-question-answer-on-next-gqaMistral (7B)#8Acc@GQA: 9.2
zero-shot-video-question-answer-on-next-qaMistral (7B)#23Accuracy: 51.1