5 ms·
Okay, it seems I should dig more on the topic. Thanks!
by jraph 3y ago
Okay, it seems I should dig more on the topic. Thanks!
- xcv123 3y agoWell as the other commenter here clarified, you're right about transformer LLMs doing next word prediction, but my point is that's only the tip of the iceberg. The complexity of the model doing the prediction is incredible. Not basic stats.