@jeffcliff @sun @RapistGoku @dictatordave it's not the model so much that matters, "we have a 10B transfomer, and 10B attention, and a 10B transformer " .. it's the training data and reinforcment that matters.