3 ms·
It's a common mistake to use ChatGPT (GPT 3.5 Turbo Model) and make statements about LLM's in general, even though GPT 4 is way ahead. Here's the result from yo
by terhechte 3y ago
It's a common mistake to use ChatGPT (GPT 3.5 Turbo Model) and make statements about LLM's in general, even though GPT 4 is way ahead. Here's the result from your prompt from GPT4:
---
Here are 20 unique letters from the Oxford English Dictionary: a, b, c, d, e, f, g, h, i, j, k, l, m, n, o, p, q, r, s, t
Now, here are 4 five-letter words using those 20 unique letters:
1. braid (b, r, a, i, d)
2. chomp (c, h, o, m, p)
3. fluke (f, l, u, k, e)
4. quest (q, u, e, s, t)
- beardyw 3y agoThat has two E and two U. Using 7 vowels will be the problem.
- aaronrobinson 3y agoI’m not really sure what your point was about a mistake, was it that I didn’t specify 3.5? Either way GPT4 fails too. It hasn’t used g as it claimed and has used e twice.