171–180 of 560 posts
When I copy-paste that error into an LLM looking for a fix, usually I get a reply in which the LLM twirls its moustache and answers in a condescending tone with a fake French accent. It is hilarious.
Earlier quoted context omitted.
But it really erodes trust. First couple of times I felt that it indeed confirmed what I though, then I became suspicious and I experimented with presenting my (clearly worse) take on things, it still said I was absolutely right, and now I just don't trust it anymore. As people here are saying, you quickly learn to not ask leading questions, just assume that its first take is pretty optimal and perhaps present it wit…
Good, because you shouldn't trust it in the first place. These systems are still wrong so often that a large amount of distrust is necessary to use them sensibly.
Earlier quoted context omitted.
You're absolutely right! Americans are a bit weird like that, most people around the world would be perfectly okay with short and to-the-point answers. Especially if those answers are coming from a machine that's just giving its best imitation of a stochastic hallucinating parrot.
Claude is very "American", just try asking it to use English English spelling instead of American English spelling; it lasts about 3~6 sentences before it goes back. Also there is only American English in the UI (like the spell checker, et al), in Spanish you get a choice of dialects, but not English.
> So... The LLM only goes into effect after 10000 "old school" if statements?
I've spent a lot of time trying to get LLM to generate things in a specific way, the biggest take away I have is, if you tell it "don't do xyz" it will always have in the back of its mind "do xyz" and any chance it gets it will take to "do xyz" When working on art projects, my trick is to specifically give all feedback constructively, carefully avoiding framing things in terms of the inverse or parts to remove.
> the biggest take away I have is, if you tell it "don't do xyz" it will always have in the back of its mind "do xyz" and any chance it gets it will take to "do xyz" You're absolutely right! This can actually extend even to things like safety guardrails. If you tell or even train an AI to not be Mecha-Hitler, you're indirectly raising the probability that it might sometimes go Mecha-Hitler. It's one of many reasons w…
Then any time the probability chains for some command approaches that locus it'll fall into it. Very much like chaotic attractors come to think of it. Makes me wonder if there's any research out there on chaos theory attractors and LLM thought patterns.
Earlier quoted context omitted.
I'm curious what Americans have to do with this, do you have any sources to back up your conjecture, or is this just prejudice?
It's common for foreigners to come to America and feel that everyone is extremely polite. Especially eastern bloc countries which tend to be very blunt and direct. I for one think that the politeness in America is one of the cultures better qualities. Does it translate into people wanting sycophantic chat bots? Maybe, but I don't know a single American that actually likes when llms act that way.
Politeness makes sense as an adaptation to low social trust. You have no way of knowing whether others will behave in mutually beneficial ways, so heavy standards of social interaction evolve to compensate and reduce risk. When it's taken to an excess, as it probably is in the U.S. (compared to most other developed countries) it just becomes grating for everyone involved. It's why public-facing workers invariably complain about the draining "emotional labor" they have to perform - a term that literally doesn't exist in most of the world!
I'm pretty sure they want it kissing people's asses because it makes users feel good and therefore more likely to use the LLM more. Versus, if it just gave a curt and unfriendly answer, most people (esp. Americans) wouldn't like to use it as much. Just a hypothesis.
If that was the case they wouldn't have so much stuff in their system card desperately trying to stop it from behaving like this: https://docs.anthropic.com/en/release-notes/system-prompts > Claude never starts its response by saying a question or idea or observation was good, great, fascinating, profound, excellent, or any other positive adjective. It skips the flattery and responds directly.
I've spent a lot of time trying to get LLM to generate things in a specific way, the biggest take away I have is, if you tell it "don't do xyz" it will always have in the back of its mind "do xyz" and any chance it gets it will take to "do xyz" When working on art projects, my trick is to specifically give all feedback constructively, carefully avoiding framing things in terms of the inverse or parts to remove.
For the 'you're right!' bit see: https://youtu.be/ZOs8U50T3l0?t=71