Sharing our experience at making LLM-based Source Code classifier
Ori Abramovsky
What to do when we train on X, inference on Y, and measure on Z
What to do what only unlabeled data is available
Yet another NLP? No No and No
3 possible pivots when a classification model training is not possible
Sub populations to take into account when sampling GitHub
From interview through task definition to evaluation
3 lean validations to make sure you're on the right path
