bjourne
5 hours ago
Probably? Reinforcement learning creates bots with specific styles. For example, ChatGPT is very fond of "typically", "unpack this", and "if you want".
Ask HN: Does Claude use 'prior' in a Bayesian sense more than English?
3 pointsposted 11 hours ago
by slakeItem id: 47970734
3 Comments
ex-aws-dude
8 hours ago
Once again a post with literally 3 points and 2 hours old is the top of /ask
Why is the HN algorithm such ass, can we talk about that?
pbkompasz
8 hours ago
Well it did have Claude both in the title and the description...