Character counts for selected authors
In training data, document size (a proxy for the number of characters in a given work) is a key indicator of how a model will behave. For example, models trained on Books3 take seven times as many instructions from Shakespeare than they do from Nobel Prize winner Alice Munro.
In training data, document size (a proxy for the number of characters in a given work) is a key indicator of how a model will behave. For example, models trained on Books3 take seven times as many instructions from Shakespeare than they do from Nobel Prize winner Alice Munro.
