4 comments

[ 0.20 ms ] story [ 22.2 ms ] thread
Do you have some test/benchmarks with the current AI engines?
I don't currently have any money to spare on the API but might create a result for GPT-4 through ChatGPT.
It looks like an interesting type of puzzle for AI. I think a blog post comparing different engines and text length would be very interesting.
Thanks and I agree. Especially because it's easily scalable to longer contexts. It's trivial to use the "framework to" run the tests but I cannot prioritize it right now so either someone will have to pick it up or I might get a time slot further on. I just wanted to publish the idea.