We design, implement, train and evaluate PRISME, a small-language-model architecture that augments a frozen
causal decoder with a read-only source bank, an ephemeral sixteen-slot latent workspace that is re-created at every
position, a shared sparse editor that rewrites one or two slots per action, a value controller trained on
counterfactual intervention gains measured by executing paired edit/no-edit branches, and a differential readout
with an exact zero-budget bypass. We reproduce the proposal's exact backbone parameter counts and its analytic
operation table, prove its structural properties (causality, bitwise identity, exact locality, bounded
displacement, one-step selection regret), and test seven pre-registered hypotheses with a complete implementation,
a fifteen-invariant test suite, an incremental inference engine and controlled, seeded comparisons against a
parameter-matched plain decoder, a dense-slot editor, a global-recurrence adapter trained with the identical
recipe, five reference routers and exhaustive one- and two-step oracles. The zero-initialised readout prescribed by
the proposal does not learn because of a scale mismatch; a scale-aware non-zero initialisation does. Sparse edits
then lower a frozen backbone's held-out loss by \vtwoprismegainbtwo{}\,\% and raise compositional exact match by
about two points, but the global-recurrence adapter matches or exceeds this at every budget. Where to write does
not matter for fixed schedules. An oracle that reads the target finds large opportunity (\routingoraclebeight{}\,\%
of loss), but \labelscalibwithin{}\,\% of the variance of intervention gains is state-level rather than action-level,
the opportunity is concentrated in high-uncertainty states where a wrong action costs more than a right one gains,
and the learned controller ranks actions within a state near chance; it is worse than no routing on held-out data.
On the locked test regimes the pre-registered primary effect is $\testprismebtwovsbzero$ points against a three-point
threshold. At 103M parameters on 98M tokens of web text, neither the sparse editor nor the global adapter moves the frozen
backbone by more than 0.1\,\% of loss and a parameter-matched plain decoder does better than both. Measured on Apple
Silicon, per-action dispatch overhead erases most of the arithmetic savings of sparsity. We report the complete negative
result on learned spatial routing, its mechanistic explanation, and the conditions under which the hypothesis could
still be rescued.