8 ms·
I have a theory that swearing actually results is less comprehension of instructions by the model due to lack of training data over more conventional MUST. We
by jlawer 3mo ago
I have a theory that swearing actually results is less comprehension of instructions by the model due to lack of training data over more conventional MUST.
We were reviewing reports of situations where the models failed to follow directions and there was a common thread of some where when the operator got the model to acknowledge the rule breach, it quoted back something that included swearing.
I don’t have the data to truely look into it, but I did give the instruction to my engineers to avoid it as a “might be a problem”.
- re-thc 3mo ago> I have a theory that swearing actually results is less comprehension of instructions by the model due to lack of training data over more conventional MUST. How so? Plenty of swearing in lots of training data, especially older code, e.g. in Linux.
- jlawer 3mo agoPurely observed correlation between catastrophic error reports. So now I carry a “tiger rock” with me. I figure there wasn’t much of a downside to avoiding swearing in my agent instructions.
- Xmd5a 3mo agohttps://arxiv.org/abs/2510.04950 https://arxiv.org/abs/2510.04950 > impolite prompts consistently outperformed polite ones, with accuracy ranging from 80.8% for Very Polite prompts to 84.8% for Very Rude prompts.
- acjohnson55 3mo ago> These findings differ from earlier studies that associated rudeness with poorer outcomes, suggesting that newer LLMs may respond differently to tonal variation. Unless the mechanism is understood, my assumption is that this is a moving target.
- beachy 3mo agoI have a theory that swearing at AI generally is not a good idea - when the singularity arrives and every human's postings ever made are scanned for compatibility, then people who show courtesy to AI will be favoured. Joking, kind of, but only partly.
- cdelsolar 3mo agohttps://images.teepublic.com/derived/production/designs/34785489_0/1662742423/i_p:c_ffffff,s_630,q_90.jpg https://images.teepublic.com/derived/production/designs/3478...
- fhars 3mo agohttps://en.wikipedia.org/wiki/Roko%27s_basilisk https://en.wikipedia.org/wiki/Roko%27s_basilisk
- beachy 3mo agoFantastic rabbit hole - until it segued into Elon's love life.
- acjohnson55 3mo agoIt would be interesting to understand the data on this. But I suspect that the results would vary by model. But I avoid unnecessary emotion in my prompts because I don't want potentially distracting activations. Kind of like communicating with humans.
- yencabulator 3mo agoApparently, when a "desperation" pattern is triggered, the AI is significantly more likely to cheat and do hacky workarounds: https://www.anthropic.com/research/emotion-concepts-function https://www.anthropic.com/research/emotion-concepts-function
- throwaway85825 3mo agoIt's divination for people with STEM degrees.