Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

How do you objectively measure if an LLM is training on something if you don't have access to its training data?


In theory the same way people are making those claims about "stolen" art, such as models that produced watermarks from Getty images or Shutterstock. Similar "watermarks" have existed in some LLM output.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: