128,000 tokens, which is about the same as a decent sized book.
Their other models can also be fine-tuned, which is kinda unbounded but also has scaling issues so presumably "a significant percentage of the training set" before diminishing returns.
128,000 tokens, which is about the same as a decent sized book.
Their other models can also be fine-tuned, which is kinda unbounded but also has scaling issues so presumably "a significant percentage of the training set" before diminishing returns.