>2. "Uncontaminated" data will be difficult to find. What with archive.org, gutenberg etc.
But recent uncontaminated data is hard to find. https://github.com/rspeer/wordfreq/blob/master/SUNSET.md
But recent uncontaminated data is hard to find. https://github.com/rspeer/wordfreq/blob/master/SUNSET.md
I really do just bail out whenever anyone uses the word slop.
>As one example, Philip Shapira reports that ChatGPT (OpenAI's popular brand of generative language model circa 2024) is obsessed with the word "delve" in a way that people never have been, and caused its overall frequency to increase by an order of magnitude.
Should run the same analysis against the word slop.
I do when people try to pass slop around as organic content.
> Should run the same analysis against the word slop.
That sounds like tautology: the more slop, the more people tend to call slop slop.