We’re looking for feedback from users working with interactive avatars. We’ve built a high-quality prompt that gives excellent, logical results when tested directly through OpenAI (e.g. in the Playground). However, when using the same prompt in the a...
