Comment by vlovich123

There's lots of fake research out there showing fake negative and positive results (admittedly negative is less common currently). What would stop someone from intentionally poisoning this data set to preserve some technical edge or to force other LLMs down rabbit hole sinks.