News for hackers. Stuff that matters.

A comment on Ask HN: How do companies like OpenAI, Perplexity fine tune rich output?.

Re:Ask HN: How do companies like OpenAI, Perplexity fine tun... by pizza on Friday July 04, 2025 @04:07AM (#44462304)
Anything with a linter means, at minimum, free verifiable rewards for RL (though whether something parses versus looks good is another story). That, plus, they have more data than anyone, and also it seems somewhat reasonable that stronger models could learn 'more' from a given instance or set of examples.
[ Parent | on HN ]
Re:Ask HN: How do companies like OpenAI, Perplexity fine tun... by agaase19 on Friday July 04, 2025 @05:51AM (#44463000)
Can you elaborate on "linter means and verifiable rewards for RL"? Is this something others would find extremely difficult to do ?
[ Parent | on HN ]
Re:Ask HN: How do companies like OpenAI, Perplexity fine tun... by holden_nelson on Saturday July 05, 2025 @12:54AM (#44470167)
They’re saying that they can use linters to check the output from a reinforcement learning model and reward it for correct output.
[ Parent | on HN ]