GPT-4o's Memory Breakthrough (Needle in a Needlestack)

Re:GPT-4o's Memory Breakthrough (Needle in a Needlestack) by personjerry on Tuesday May 14, 2024 @04:46PM (#40359975)

That's great to hear. My biggest issue with GPT-4.0 was that as the conversation got longer, the quality diminished (especially relevant for coding projects)

I wonder if it'll be better now. Will test today.

[ Parent | on HN ]

Re:GPT-4o's Memory Breakthrough (Needle in a Needlestack) by throwthrowuknow on Tuesday May 14, 2024 @06:07PM (#40360698)

That’s been my experience so far. My current conversations are crazy long compared to any of my gpt4 convos which I had to frequently copy context from and start over in a new chat

[ Parent | on HN ]

Re:GPT-4o's Memory Breakthrough (Needle in a Needlestack) by sftombu on Tuesday May 14, 2024 @04:50PM (#40360005)

I had the same experience. With a 16k prompt, Turbo was nearly flawless. But it wasn't very good at 32k and not usable at 100+. You have to repeat information to get good results with longer prompts

[ Parent | on HN ]