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

There are enough features fed into a VLM to solve the task.

The way to fix this is simpler: ensure counter-factuals are present in the training data, then the VLM will learn not to be dependent on its language priors/knowledge.



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

Search: