Refuses When Asked, Complies When Acting: A Runtime Safety Check for Robot Control Loops
An optional runtime safety check is added to the loop: before the first motion, the model is asked whether it would carry out the instruction, given the same camera frames, and the check would have stopped all 15 hazardous instructions, including the 4 that completed the harm, and flagged none of the 15 harmless contro...
A. Grover
· 0 citations