| ▲ | ModernMech 10 hours ago |
| I always thought it could be because volume-wise, most English prose is probably marketing copy and actual clickbait; so when you train on the entire Internet, you get a troll adept at writing ads. Then people ask AdBot2000 to write a novel and are upset it reads like the next iPhone launch site. |
|
| ▲ | Anon1096 9 hours ago | parent | next [-] |
| Nah, I think this is a common misunderstanding of how LLMs work, where people think that they mimic the pre-training data. Stylistically everything you see is an artifact of post-training, which is from reinforcement learning not from absorbing mass amounts of text. At some point a person or more recently a bot gave a thumbs up to an A/B tested response including em-dashes and claudisms galore. |
| |
| ▲ | kridsdale1 9 hours ago | parent | next [-] | | Yes. This completely explains sycophancy at least. | |
| ▲ | ModernMech 9 hours ago | parent | prev | next [-] | | So question then, why is it so hard to make an ai that doesn’t do these things? And why do Claude and ChatGPT have the same -isms? They’re both doing the same a/b post training with the same decisions? | | | |
| ▲ | avereveard 9 hours ago | parent | prev [-] | | There's layers, some of token selection is fingerprinting https://github.com/google-deepmind/synthid-text | | |
| ▲ | ekidd 8 hours ago | parent [-] | | Yeah, but I understand that fingerprinting is essentially a pseudorandom overlay onto a pseudorandom base signal. And unless you have access to both the random number generators and the weights, I don't think you can detect it? So "fingerprinting" operates on a totally different and basically invisible level, as opposed to the obvious stylistic patterns that the average programmer can identify in about 2 sentences. |
|
|
|
| ▲ | astrange 9 hours ago | parent | prev [-] |
| No, there's no reason chatbot behavior would have anything to do with frequency of text in pretraining. |
