This week's Paper Club is focused on data. As models have scaled across tasks and languages, old assumptions about training and evaluation are starting to break.

So we gathered three domain experts to break down the challenges and frontiers of training data, benchmarks, and