Id love to learn what prompts you’re using to do that. Explanations one at a time. Thats something I’ve thought would be helpful before but didn’t know how to achieve it.
For me this works: "One thing at the time. This is a conversation, not a lecture, try to balance the amount of words you write versus the amount of words that I write. Don't overwhelm me with 10 pages of prose where a single sentence would suffice, brevity is a virtue, not a defect." I've had to tweak it a couple of times, probably because of different versions of ChatGPT but it is usually a variation on this that will do the trick. Besides being far more interactive it takes the frustration down quite a bit.
It slows down to the point I find it unusable so I ask it to summarize as one cut-and-pastable block, copy the block, open a new session and kill the old one. This usually happens after a few hours, probably because 'thinking tokens' even if they are not displayed crowd the context window.
Nothing fancy. What works for me is trying to lead the conversation: asking for definitions one at a time, asking how they differ from its previous answers and pointing it out when it conflates its own answer. I generally feel like you can't let it dictate the pace, it's not very good at that yet.
I used to edit my prompts and undo messages if it misunderstood something but that's been failing me recently.
One of the things that I learned the hard way is to not to combine requests, but fork chats _a lot_ and ask singular questions. Also start new sessions with either explicitly created handovers or even just explaining current state. The more compactions I see the less and less trust I have in it’s current understanding of what we’re discussing
That's interesting. The "best" current models, Fable and especially GPT 5.6, are also lying liars that lie all the time. Seems like we're going the wrong way on hallucination.
I've been seeing more "grounding" messages in the thinking log of Claude models. I guess that's "do a web search, read the docs", probably designed to mitigate hallucinations.
I am surprised by how much the really giant models hallucinate, though. My vague feeling was that little models hallucinate a lot because they just don't know anything (the world's knowledge simply does not fit in a few GB) and don't know how to say, "I don't know". But, the big models kinda do know everything, and yet, here we are, they're still making shit up all the time.
> I am surprised by how much the really giant models hallucinate, though. My vague feeling was that little models hallucinate a lot because they just don't know anything (the world's knowledge simply does not fit in a few GB) and don't know how to say, "I don't know". But, the big models kinda do know everything, and yet, here we are, they're still making shit up all the time.
In my mostly unscientific RAG experiments I found that larger models were more likely to hallucinate in a RAG setting. I think that it's because they have more world knowledge. As an example, when asking about safety legislation in Ireland, Claude got hooked on the notion that OSHA was involved, which clearly is not true, while the smaller models just said that they couldn't find any information (which was the correct answer).
One of my tests to see if "we are there", it's to ask a list of mayors from my city. Sounds silly, but most models just create random mayors that never existed.
Without search, of course. In theory this should be easy, because clearly the Wikipedia data it's in the model.
reply