News for hackers. Stuff that matters.
Re:Show HN: Compress GPT-4 Prompts by staunton on Friday April 07, 2023 @08:46PM (#35489147)
Indeed, LLM's seem to be much worse at introspection than humans. I wonder what would happen if one used reinforcement learning to train into it the ability to correctly predict and reason about it's capabilities and behavior.
Re:Show HN: Compress GPT-4 Prompts by MacsHeadroom on Saturday April 08, 2023 @04:13AM (#35491540)
Then you would have designed https://github.com/Torantulino/Auto-GPT

(Uses recurrent langchain loops for introspection and learning about itself and its capabilities as they grow + vector databases like Pinecone for long term memory)