← Back to context Comment by anitil 1 year ago Oh this is like 'Reflections on Trusting Trust' for the AI age! 1 comment anitil Reply kibwen 1 year ago With the caveat that the attack described in RoTT has a relatively straightforward mitigation, and this doesn't. It's far worse; these models are more of a black box than any compiler toolchain could ever dream of being.
kibwen 1 year ago With the caveat that the attack described in RoTT has a relatively straightforward mitigation, and this doesn't. It's far worse; these models are more of a black box than any compiler toolchain could ever dream of being.
With the caveat that the attack described in RoTT has a relatively straightforward mitigation, and this doesn't. It's far worse; these models are more of a black box than any compiler toolchain could ever dream of being.