elsatch

Great work!

How does this benchmark translate to other languages beyond English? Reading the paper, I understand your decision to focus on a single language first, but to me this feels like the MLEEB (Massive Legal English Embeddings Benchmark).

Are you planning for a MMLEB?

Very interesting idea! I can't way to apply this to other Spanish datasets!

I think @singhsidhukuldeep might be talking about some kind of OctoLand Architecture :)

I think this might be covered in the Kraken LoRA folder, that was also recently released: https://github.com/cognitivecomputations/kraken/tree/main/Kraken-LoRA

cc @fernandofernandes @DavidGF @Crystalcareai @ehartford

Maybe @alexsherstinsky can solve the question about the multiple LoRA adapters :)