On: The genome diversity of major crops tells the stor...

[pests]

That's assuming a pretained model where you are just stuffing the genome data into the prompt.

You could fine-tune or train your own model on the data and then design the prompt / query interface to give you interesting results.

[Myrmornis → pests]

Yes, true. It's not clear that next word prediction is the best model for such an exercise, but yes, they do seem rather successful.

[pests → Myrmornis]

Everyone always calls it "next word prediction" but that's also a simplification.

If you go back to the original Transformer paper, the goal was to translate a document from one language to another. In all prompt systems the model is only given "past" tokens (as it's generating new tokens in real time) but in that original paper the LLM can use backwards and forward context to determine the translation.

Just saying the architecture of how a model is trained and how it outputs tokens is less limited than you think.