In Search of Goodness: Large Scale Benchmarking of Goodness Functions for the Forward-Forward Algorithm