Concrete benchmark proposals for how to detect mode-collapse and AI slop and ChatGPTese, and why I think this might be increasingly important for AI safety, to avoid ‘whimper’ or ‘em hell’ kinds of existential risk: https://gwern.net/creative-benchmark EDIT: resubmitted as linkpost.
I observed an effect of “chatification”: my LLM-powered mind model tell me stories about my life and now I am not sure what is my original style of telling such stories.
Concrete benchmark proposals for how to detect mode-collapse and AI slop and ChatGPTese, and why I think this might be increasingly important for AI safety, to avoid ‘whimper’ or ‘em hell’ kinds of existential risk: https://gwern.net/creative-benchmark EDIT: resubmitted as linkpost.
People inexplicably seem to favor extremely bad leaders-->people seem to inexplicably favor bad AIs.
I observed an effect of “chatification”: my LLM-powered mind model tell me stories about my life and now I am not sure what is my original style of telling such stories.