News for hackers. Stuff that matters.

A comment on MemGPT – LLMs with self-editing memory for unbounded context. Parent comment

Re:MemGPT – LLMs with self-editing memory for unbounded context by JCharante on Monday October 16, 2023 @01:32PM (#37903283)
To me it just feels like they’re trimming the min amount of oldest tokens in the conversation to stay under the token limit. Conversations don’t degrade in a way that feels like it has medium term memory.
[ Parent | on HN ]

There are no comments on this comment yet.