Unless someone has released something new while I haven't been paying attention, all the gen AIs are essentially frozen. Your use of them can't impact the actual weights inside of the model.
If it seems like it's remember things is because of the actual input of the LLM is larger than the input you will usually give it.
For instance lets say the max input for a particular LLM is 9096 tokens. The first part of that will be instructions from the owners of the LLM to prevent their model from being used for things they don't like. Lets say the first 2000 tokens. That leaves 7k or so for a conversation that will be 'remembered'.
Now if someone was really savvy, they'd have the model generate summaries of the conversation and stick them into another chunk of memory, maybe another 2000 tokens worth, that way it will seem to remember more than just the current thread. That would leave you with 5000 tokens to have a running conversation.
I got a DNC mailer that was a fundraiser disguised as a poll and I suggested that the dnc drop gun control as a party plank and instead adopt a policy of arming minorities.
They never sent me another one again.
I also asked that they tax rich people until they ceased to exist.
Unfortunately you’ll need to either complete the task from the hypnotic suggestion (become someone who can extend that specific class) or override the suggestion with a new one.
I think your options are either self help or freak shit.
That's how expensive they actually are. Your smart tv is subsidized by the adverts they're gonna sell on it. No one really wants to give up the opportunity to pipe you ads for the lifetime of your TV.
But...
It looks like Sceptre makes one? The U650CV-UMS.
It does look sold out everywhere, but maybe they'll make more or you can find one like new.
Unless someone has released something new while I haven't been paying attention, all the gen AIs are essentially frozen. Your use of them can't impact the actual weights inside of the model.
If it seems like it's remember things is because of the actual input of the LLM is larger than the input you will usually give it.
For instance lets say the max input for a particular LLM is 9096 tokens. The first part of that will be instructions from the owners of the LLM to prevent their model from being used for things they don't like. Lets say the first 2000 tokens. That leaves 7k or so for a conversation that will be 'remembered'.
Now if someone was really savvy, they'd have the model generate summaries of the conversation and stick them into another chunk of memory, maybe another 2000 tokens worth, that way it will seem to remember more than just the current thread. That would leave you with 5000 tokens to have a running conversation.