part 3 in LLM interface exploration—combine words to generate new concepts
Thread
This is cool! Which model are you using? Looks like there’s almost no latency
Oh I didn't notice it's sped up and even considering that, it looks super smooth. Nice work!