3 ms·
Yeah I hate that bug! The thing where models suggest that they will take your feedback into account, when you know that they can't do that.
by simonw 2y ago
Yeah I hate that bug! The thing where models suggest that they will take your feedback into account, when you know that they can't do that.
- sebzim4500 2y agoIt's possible that OpenAI scrapes people's chat history for cases where that happens in order to improve their fine tuning data, in which case it isn't a total lie.
- outofpaper 2y agoThey say that will do this so long as users don't opt out.
- radicality 2y agoIsn’t that kind of how the “ChatGPT memory” feature works? I’ve recently seen it tell me that it’s updating memory, and whatever I said does appear under the “Memory” in the settings. I’m not familiar though with how the Memory works, ie whether it’s using up context length in every chat or doing something else.
- simonw 2y agoYeah, memory works by injecting everything it has "remembered" as part of the system prompt at the beginning of each new chat session: https://simonwillison.net/2024/Feb/14/memory-and-new-controls-for-chatgpt/ https://simonwillison.net/2024/Feb/14/memory-and-new-control...
- joquarky 2y agoI'm curious how it will avoid filling up the context window as more memories are added.
- refulgentis 2y agoCount the tokens, RAG em down to max count you're willing to pay for
- hackable_sand 2y agoVector compression. "Summarize memories at timestamps: [i, j, k, ...n]" etc. etc. until you have a nice scrapbook of memories. Then you gotta prune low-traffic, low-value vector spaces by whatever heuristic you (or an automated system) deems important.
- mewpmewp2 2y agoIsn't that also what people always say out of politeness?