I don't think it is reasonable to ask for specific model versions in AI disclosures. If you work on a paper for 2 years you will probably use 20+ different models just by using the most recent one. Should we also report what ever Gemini model Google uses for the AI search overview?

Show Post

Internal within, external between. Needs a multilevel model

Show Post

I have some metaphysical questions about this subject line

Show Post

Schrödinger’s newsletter

Show Post

Essentially, it's just a very long table listing all the studies and summarising their findings. I started working on this study before LLMs became popular. I wouldn't do it again by hand.

Show Post