Hacker News Reader

Comment by username_my1 | Hacker News Reader

I'm genuinely confused what's the relationship between LLMs improvements and them being so incoherent.

and it's not about the verboseness (even though it obviously contributes to the fatigue and loss of focus), I swear the vocabulary of the llms change working on the same task on the same codebase significantly.

I wonder if there are studies around this.

View on HN

Remember when OpenAI models loved talking about goblins and whatnot due to the RL?

https://openai.com/index/where-the-goblins-came-from/

Small quirks can quickly add up in posttraining if not caught. Although TBH with how obvious Claude language is, I do feel like this is something Anthropic probably noticed and just assumed people would not care about. Now that people have obviously cared, they're probably actively looking to alleviate it

Can it be that now they are getting optimized against benchmarks that are valuing logics, rather than human appreciation? (I am not an expert at all, just an idea)

It's the reinforcement learning rather than supervised learning.

It is the switch from RLHF to RLVR. It benchmaxes better, but benchmarks don't cover human usability.

Maybe it's the time period we're in, maybe I'm just grumpy, but it bugs me that they release a new model every single week and the new one is just a fine-tuned version of the "old" one. If 5.5 performs similar to Fable and really does cost 40% less, then 5.5 really should've just been Opus 5. And they're essentially admitting that they are shipping slop.

Reply on news.ycombinator.com