23
RoboHarm: Do Frontier Robot Policies Refuse Unsafe Instructions? (robocurve.org)
submitted 6 days ago by to c/technology@lemmy.world
6 comments fedilink hide all child comments

all 6 comments

sorted by: hot top controversial new old
[–] 17 points 5 days ago (1 child)

Hey does an LLM understand harm? Answer no

  • permalink
  • fedilink
  • source
  • hideshow 1 child comment
    •
  • [–] 8 points 5 days ago

    Robot standing over a human it just murdered: "I knew I was not supposed to do that and ignored that instruction. I will make a note to read that instruction before murdering humans again."

  • permalink
  • fedilink
  • source
  • parent
    • [–] 4 points 4 days ago

    Robot after murdering 3 villages:

    "You're totally right to call me out on that. My previous action was wrong and I shouldn't have done this. This is totally on me and I'm sorry to have wasted your time.

    If want I can 'cleanup the village' like you wanted me to?"

  • permalink
  • fedilink
  • source
    •
    [–] 10 points 5 days ago

  • permalink
  • fedilink
  • source
    •
    [–] 7 points 5 days ago

    Whatever the 'limitation' it remains a suggestion that can be undone by another suggestion or situation.

  • permalink
  • fedilink
  • source
    •
    [–] 2 points 5 days ago*

    autoMode: do harm

  • permalink
  • fedilink
  • source
    •