I built KillSwitch, an esoteric language designed to confound LLMs and coding agents, and benchmarked the major models against it.
  • freely1333@reddthat.comEnglish
    10·
    16 hours ago

    If you modeled your language around bio weapon, red teaming, and super racist language an LLM might actually refuse.

    • Windex007@lemmy.world
      1·
      1 hour ago

      Rot13 every keyword in and out before the agent sees it (solution for anthropic models)

      Keyword replace things to be even more offensive to make grok operate at peak efficiency.

    • phlegmy@sh.itjust.works
      8·
      13 hours ago

      Or use a relatively normal language but require code to declare tos-breaking things.

      #error, line 47: ‘return’ called before declaring trump a pedophile.

      #error, line 59: ‘while’ used in method before declaring intent to bomb a data centre.