3 ms·
llama2 was kinda weird. I work on postgres internals for a living, I asked it information about a postgres internal function - it responded that it was unethica
by lumost 3y ago
llama2 was kinda weird. I work on postgres internals for a living, I asked it information about a postgres internal function - it responded that it was unethical to ask about private/internal postgres functions. It then proceeded to expound on the perceived ethics of the situation.
- mhh__ 3y agoIn the very early days of llama2 at least it refused to write poetry on the basis that it was improper to do so
- andrelaszlo 3y agoLLM poetry is still quite vogonic, so perhaps it was an ethically sound refusal?
- gryn 3y agoThat not just a llama specific thing most llama do it. I've had that with openAI and gemini too. It just seems like their way to virtue signal while covering their ass legally.
- ranyume 3y agoI believe that's only half the issue. I'm suspecting companies don't actually know how to make an "uncensored" model (a model that acts like a tool, not refusing anything). The way the refusals are phrased and how they respond if you press them to explain themselves is just too quirky to me. Even though the models respond with "I'm a model" in my tests they get pretty emotional in their refusals. "I feel uncomfortable with that". This makes me think there's something wrong in their training process that's making these refusals. Or at least it's making refusals less predictable (false positives).