@ahmadmanga's thread
73
@ahmadmanga
· 2Y agoI Asked ChatGPT: Is it beneficial to train a LLM on other LLMs prompted responses?
Answer:
Training an LLM on responses from other LLMs can increase data volume and speed up training but risks inheriting errors, biases, and reducing originality. Human-generated data is typically more reliable for building robust models.
Expanded answer below in comments:
No replies made yet. Would you like to be the one to do so?
