Yes, but I highly doubt they would increase sparsity much vs the chinese models. That's how you get Llama 4. Pretty much every major lab settled on ~3-5% spar

Yes, but I highly doubt they would increase sparsity much vs the chinese models. That's how you get Llama 4. Pretty much every major lab settled on ~3-5% sparsity for a reason.