Comment Comment by minxomat The whole point of the LLaMa paper is that large models are undertrained and oversized. parent minxomat 3 years ago
