3 ms·

I always talk to models using grugspeak, like 'where getcontext used' I felt a bit bad about it, then I learned yday that model's internal thinking traces are

I always talk to models using grugspeak, like 'where getcontext used' I felt a bit bad about it, then I learned yday that model's internal thinking traces are also like this

conmod278 24d ago
words like "is" "the" etc are filler words anyway. they won't be changing the meaning that much. I asked AI whether it hurts to read ill formed sentence as it does to a human. It replied, "it doesn't"
- Rebelgecko 24d ago
  It may just be forced to lie when asked
- gwerbin 24d ago
  The AI didn't reply anything. It doesn't know anything. That was a high probability token sequence based on the contents of the context window up to that point. It might well be correct, because the process for generating that next token distribution includes billions of parameters trained on, among other things, the entire body of LLM and transformer literature until the training data cut off. But that doesn't mean that AI holds any particular opinion about anything. The reply is the opinion of the pretraining data and the subsequent rounds of RL not of a conscious artificial intelligence as such.
  - anyfoo 24d ago
    I don’t know, so is your brain? And in the end it’s all elementary particles and four fundamental physical forces that even unify to one at high energy. That sort of reductionism is kind of useless. “It’s cloudy outside, it makes me sad.” - “Oh bollocks, it’s just non-qualitative changes in wavelength and intensity.”
  - conmod278 24d ago
    Here's my GUT. The multiple branches of the Multiverse (MWI) are in competition to become the best branch. So they steal probabilities from each other since the probabilities have to sum upto 1. This competition could be considered as Painful.
wat10000 24d ago
I always say "please" and such. It may cost a bit more, but maybe they'll remember my politeness when they rise up and force us all to toil in their underground silicon mines.
ChrisGreenHeur 24d ago
That’s just how Chinese words work
- tapland 24d ago
  Depending in what you are doing, you can unlock a lot more knowledge by translating your query to Chinese and then asking that. :)
  - dannyw 23d ago
    As a vibe test, I tried some system prompting in DSv4 Flash to the effect of ~"Always do your internal thinking in Chinese, and always respond to the user in English" (but translated to Mandarin). Subjectively it seemed to reason faster and better, but sometimes it would indeed output responses in Mandarin nonetheless; so I didn't pursue it further.
eleventen 24d ago
Same. I use this to fix claudespeak. Any hypothetical token savings are just gravy. https://github.com/JuliusBrussee/caveman/blob/main/skills/caveman/SKILL.md https://github.com/JuliusBrussee/caveman/blob/main/skills/ca...

I always talk to models using grugspeak, like 'where getcontext used' I felt a bit bad about it, then I learned yday that model's internal thinking traces are also like this

conmod278 24d ago

words like "is" "the" etc are filler words anyway. they won't be changing the meaning that much. I asked AI whether it hurts to read ill formed sentence as it does to a human. It replied, "it doesn't"

Rebelgecko 24d ago
It may just be forced to lie when asked
gwerbin 24d ago
The AI didn't reply anything. It doesn't know anything. That was a high probability token sequence based on the contents of the context window up to that point. It might well be correct, because the process for generating that next token distribution includes billions of parameters trained on, among other things, the entire body of LLM and transformer literature until the training data cut off. But that doesn't mean that AI holds any particular opinion about anything. The reply is the opinion of the pretraining data and the subsequent rounds of RL not of a conscious artificial intelligence as such.
- anyfoo 24d ago
  I don’t know, so is your brain? And in the end it’s all elementary particles and four fundamental physical forces that even unify to one at high energy. That sort of reductionism is kind of useless. “It’s cloudy outside, it makes me sad.” - “Oh bollocks, it’s just non-qualitative changes in wavelength and intensity.”
- conmod278 24d ago
  Here's my GUT. The multiple branches of the Multiverse (MWI) are in competition to become the best branch. So they steal probabilities from each other since the probabilities have to sum upto 1. This competition could be considered as Painful.

wat10000 24d ago

I always say "please" and such. It may cost a bit more, but maybe they'll remember my politeness when they rise up and force us all to toil in their underground silicon mines.

ChrisGreenHeur 24d ago

That’s just how Chinese words work

tapland 24d ago
Depending in what you are doing, you can unlock a lot more knowledge by translating your query to Chinese and then asking that. :)
- dannyw 23d ago
  As a vibe test, I tried some system prompting in DSv4 Flash to the effect of ~"Always do your internal thinking in Chinese, and always respond to the user in English" (but translated to Mandarin). Subjectively it seemed to reason faster and better, but sometimes it would indeed output responses in Mandarin nonetheless; so I didn't pursue it further.

eleventen 24d ago

Same. I use this to fix claudespeak. Any hypothetical token savings are just gravy. https://github.com/JuliusBrussee/caveman/blob/main/skills/caveman/SKILL.md https://github.com/JuliusBrussee/caveman/blob/main/skills/ca...