3 ms·
Instead of complaining, why not show the benchmarks? Like: first it scored 83, now it scores only 42 (or whatever).
by amelius 3y ago
Instead of complaining, why not show the benchmarks?
Like: first it scored 83, now it scores only 42 (or whatever).
- ShamelessC 3y agothe commenters of HN are simply so smart that their hunch clearly holds more weight than scientific rigor.
- sebzim4500 3y agoWhat benchmarks? We have no way to go back in time and run the old ChatGPT accessble models though benchmarks. And to my knowledge, no one has copy pasted an entire benchmark into the ui in order to run it, they just use the API.