Birdwatch Note Rating
2023-08-15 17:57:13 UTC - SOMEWHAT_HELPFUL
Rated by Participant: EB7ED4475B01596A3502606752145DDDD3836718FFD674AAD61A916D4AA266E9
Participant Details
Original Note:
This article does not disclose that the study it cites uses the much less capable GPT-3.5-turbo model, rather than the more advanced GPT-4. Additionally, the study used the API rather than the website and so intentionally chose to use a worse model. https://arxiv.org/pdf/2308.02312.pdf (model on bottom of page 3)
All Note Details