[Question] What organizations other than Conjecture have (esp. public) info-hazard policies?

I believe Anthropic has said they won’t publish capabilities research?
OpenAI seems to be sort of doing the same (although no policy AFAIK).
I heard FHI was developing one way back when...
I think MIRI sort of does as well (default to not publishing, IIRC?)

No comments.