GPT-4 - Hackdot

Re:GPT-4 by JimDabell on Tuesday March 14, 2023 @02:29PM (#35156288)

You can already use an LLM to train a smaller, more efficient LLM without significant loss in results.

[ Parent | on HN ]

Re:GPT-4 by canoebuilder on Tuesday March 14, 2023 @02:52PM (#35156628)

Do you mean the output of a LLM as the training data for the new model? What is the specification for the prompts that generate the training data?

Any links with more info?

[ Parent | on HN ]

Re:GPT-4 by luxcem on Tuesday March 14, 2023 @03:18PM (#35156967)

There were just an article submitted few days ago about Alpaca, a LLM trained on GPT prompts: https://news.ycombinator.com/item?id=35136624

[ Parent | on HN ]

Re:GPT-4 by canoebuilder on Wednesday March 15, 2023 @12:56AM (#35164061)

Thanks!

[ Parent | on HN ]