4 ms·
AI labs should put some incentives in their RL to make their models write shorter code so it's easier to check
by sometimelurker 5mo ago
AI labs should put some incentives in their RL to make their models write shorter code so it's easier to check
- georgeven 5mo agoThe issue then becomes it writes insane one liners that are impossible to read. Would be difficult to set up an RL environment where the goal is correctness and readability.
- sometimelurker 5mo agoyou could have an AI judge look for correct and readable code