Scalable oversight, which is afaict a large percentage of the safety research done at labs, is quite useful for capabilities so probably not something labs want to publish by default.
That makes sense. Dual-use work is definitely hard to deal with, and I don’t have any good ideas on how to diffuse the safety-relevant parts across labs with acceptably low risk of spreading capabilities-related knowledge (if it’s even possible to separate these parts). Perhaps a trusted intermediary could help evaluate which details are safe to share? Do you have any ideas on how to approach publishing dual-use safety research responsibly (if at all)?
Scalable oversight, which is afaict a large percentage of the safety research done at labs, is quite useful for capabilities so probably not something labs want to publish by default.
That makes sense. Dual-use work is definitely hard to deal with, and I don’t have any good ideas on how to diffuse the safety-relevant parts across labs with acceptably low risk of spreading capabilities-related knowledge (if it’s even possible to separate these parts). Perhaps a trusted intermediary could help evaluate which details are safe to share? Do you have any ideas on how to approach publishing dual-use safety research responsibly (if at all)?