67 1 week ago

Like a little baby with a small brain.

0.0001m 0.00077m 0.001m 0.26m 1m
429eb78516da · 183B
MIT
small-minded:768 - exactly 768 parameters. d_model 2, 1 layer, ffn 38, 259-token SPM byte vocab.
Cannot read input: at this size every prompt character tokenized to <unk> anyway.