> Then, by giving them context or by post-training, you can make them sample non-average parts of the distribution they learned.
How do you derive that something is "below average" or "average" or "above average"?
How do you derive that something is "below average" or "average" or "above average"?
In the case of real world LLMs and post-training, what is above average is defined roughly as: labeled good by expert humans, and scoring high on RL environments related to coding like debugging, passing tests, or running efficiently and verifiably correctly.
One technique is RLHF: have an human expert assess it.
Like a short example is easier to grade, but not in the same ballpark as a whole codebase.
How do you? I mean, that was your point basis.