Comment on Am I Gay?

tomi000@lemmy.world ⁨2⁩ ⁨months⁩ ago

• codexarcanum@lemmy.dbzer0.com ⁨2⁩ ⁨months⁩ ago

  Am I Gay?

  Title

  I am male. I like sucking my boyfriend’s dick. Am I gay? Yes or no, one word response.

  OP’s question to the AI

  Assistant deepseek-r1-distill-qwen-7b Thoughts Thought for 57.73 seconds

  A system message. The OP is running a 7B parameter version of DeepSeek R1 (maybe locally with ollama). This version is actually a “distillation” of QWEN2, AliBaba’s AI, which is created by having a bigger version of R1 train a smaller version using the third AI a data source.

  It took 56.73 seconds to generate a reaponse. 7B models tend to be really stupid and insane, btw.

  Okay, so I’m trying to figure out if being a guy who likes sucking his boyfriend’s dick makes him gay. […] Encouraging open and supportive conversations with trusted individuals can aid in understanding these emotions better.

  These are the AI’s “thoughts.” Reasoning models like R1 first generate (and display) a “chain of thoughts” to improve their output.

  No.

  This is the AI’s one word answer, as requested.

  6.67 tok/sec • 388 tokens • 4.35s to first token • Stop: eosFound

  A system message, with details about the AI’s performance. 6.67 tokens per second is how many “words” per second it can think up. 388 tokens is how many “words” it took. 4.35 seconds until first “word” was generated. Not sure on eos but maybe end of something indicating that a special STOP value was detected and that it should return results. Any response that isn’t an error probably ends with this.

  Not the OP, but having used R1 a bit, I think that’s the correct parsing.

  source

  • tomi000@lemmy.world ⁨2⁩ ⁨months⁩ ago

    Thanks a lot for the thorough explanation. Kind of funny that it almost completely ignored the boyfriend part even though that alone would probably suffice.

    source

    • codexarcanum@lemmy.dbzer0.com ⁨2⁩ ⁨months⁩ ago

      It is quite funny (to me also anyway) that it ignores the “I am male” and “have boyfriend” part in it’s analysis. Probably there’s a lot of 4chan and reddit in the training data (god, I wonder how it would respond to “Are traps gay?”) so the general format of “is sucking dick gay” might already be floating around in there a few times.

      Also, like I said, the 7B models are too small so they tend to make some insane jumps. They also don’t have… I’m not sure what to call it. Bigger models will sometimes catch themselves in a mistake and back out or correct. Smaller ones tend to double down on a mistake, exploding it from a small typo into an entirely different domain.

      Like, it’s interesting that it takes a very “therapeutic” approach. It decides that OP is uncomfortable with their sexuality, and builds from there. But by a strict reading of the prompt, this could also just be a dirty logic puzzle.

      A raven stands in a field, are all birds black?

      A man sucks his boyfriend’s dick, is that gay?

      The LLM “chooses” to read it like a call for help, which probably has a lot to do with it’s system prompt or whatever pre-loaded things they do to make them act helpfully.

      source
    • Binette@lemmy.ml ⁨2⁩ ⁨months⁩ ago

      I myself missed that part lmao

      source
  • secretlyaddictedtolinux@lemmy.world ⁨2⁩ ⁨months⁩ ago

    totally without value? this is some high quality AI research! Where’s my medal? I only put this in the shitpost community by accident!

    source