Brendan Dolan-Gavitt @moyix.net · Apr 1 Announcing CheatGPT, a revolutionary model that achieves SoTA on HumanEval! It's incredibly sample-efficient – just ONE training sample – and *tiny*, fitting on your Casio wristwatch!
6 likes 1 replies
?
Replies
This makes me wonder if a corpus of similar questions that would intentionally be hard to overfit against is possible