> For LLMs, “open source” is a bit of a misnomer.
This is an understatement: the “source” for an LLM or similar deep learned network is not the weights, but the training data and the scripts required to train it from scratch.
This is an understatement: the “source” for an LLM or similar deep learned network is not the weights, but the training data and the scripts required to train it from scratch.
No comments yet.