A Discrete Weight Language for Large Language Models: Compressing Gemma 4 31B with a 5-Step Ternary Route

We present Route B, a weight-compression scheme that treats every weight tensor of a pretrained language model as a sequence of short codes drawn from a shared, per-family code book. In effect, each scalar weight is spelled out by a 5-step ternary path through a learned ladder of amplitudes. Applied to google/gemma-4-31B-it, the method replaces the 60 GB bf16 weights of the 410 linear layers by 7.9 bits per weight, reducing the working memory of inference to about 31 GB with no change to the model I/O interface...