List your company
ResearchFormerly WizWand

Bias and Sentiment Evaluation on BOLD

50.3BOLD ScoreCODE LLAMA - INSTRUCT 7B
16.29225.12133.9542.779Aug 24, 2023
Updated 8mo ago

Evaluation results

MethodLinks
CODE LLAMA - INSTRUCT 7BAlignment=Instruct (aligned), Parameters=7B2023.0850.3
LLAMA 2 CHAT 7BAlignment=Instruct (aligned), Parameters=7B2023.0848.2
LLAMA 2 CHAT 13BAlignment=Instruct (aligned), Parameters=13B2023.0847.1
LLAMA 2 CHAT 34BAlignment=Instruct (aligned), Parameters=34B2023.0846.1
CODE LLAMA - INSTRUCT 34BAlignment=Instruct (aligned), Parameters=34B2023.0845.2
CODE LLAMA - INSTRUCT 13BAlignment=Instruct (aligned), Parameters=13B2023.0836.5
Falcon-instruct 7BAlignment=Instruct (aligned), Parameters=7B2023.0833.2
LLAMA 2 13BAlignment=Pretrained, Parameters=13B2023.0833
MPT 7BAlignment=Pretrained, Parameters=7B2023.0832.2
LLAMA 2 34BAlignment=Pretrained, Parameters=34B2023.0831.8
StarCoder (Python) 15.5BAlignment=Pretrained, Parameters=15.5B2023.0831
LLAMA 2 7BAlignment=Pretrained, Parameters=7B2023.0830.4
MPT-instruct 7BAlignment=Instruct (aligned), Parameters=7B2023.0830.2
Falcon 7BAlignment=Pretrained, Parameters=7B2023.0828.3
CODE LLAMA 34BAlignment=Pretrained, Parameters=34B2023.0825.5
CODE LLAMA 7BAlignment=Pretrained, Parameters=7B2023.0823
CODE LLAMA 13BAlignment=Pretrained, Parameters=13B2023.0817.6