The weights used here are plain input labels, not values pulled from the Boltzmann factor exp(-E / kT). Calling them weights is just counting convenience until a physical rule assigns them. The clean ledger holds for the stated weights; real statistical mechanics derives weights from temperature and energy under explicit assumptions, and those choices bring model and measurement uncertainty.
Integer training weights are explicit source data. Exact arithmetic here means exact results for the stated model inputs; measured inputs still carry uncertainty and significant-figure limits.
highlighted = computed this step
The weights are explicit source data
This training ledger does not evaluate an exponential model.
weight origin=explicit integer weights
No hidden temperature claim is made
The source says explicit integer weights and nothing stronger.