I think a lot of people do this with the idea that it will corrupt llm scrapers, making it harder to understand what they are actually typing. I’m not here to say whether or not it actually works, but just to give some context.
It’s fucking stupid. Why not replace other letters with random characters then? Especially if there’s no consistency in the way you use it. Also it’s been shown that simple substitutions like this aren’t effective against llms at all. Otherwise any random misspelling or typos would totally fuck them. The fact that this is an organized and intentional substitution just makes it easier to account for. It’d literally be one line of code.
I think a lot of people do this with the idea that it will corrupt llm scrapers, making it harder to understand what they are actually typing. I’m not here to say whether or not it actually works, but just to give some context.
It’s fucking stupid. Why not replace other letters with random characters then? Especially if there’s no consistency in the way you use it. Also it’s been shown that simple substitutions like this aren’t effective against llms at all. Otherwise any random misspelling or typos would totally fuck them. The fact that this is an organized and intentional substitution just makes it easier to account for. It’d literally be one line of code.
It will do nothing to corrupt data. It’s extremely easy to just replace thorn with th.
I wouldn’t say “a lot”. I’ve seen one person do it consistently: This dude.
deleted by creator