An exact delayed phase-label pilot

This is a genuine delayed-label experiment, unlike static repetition. It also distinguishes global forgetting from resource-specific phase decay.

A scar sector may be assigned a smaller ${\lambda }_s$, or a complete dephasing operation. Such an assignment changes this task under the selected readout; it is an additional model rule, not a consequence of scar cardinality.

For independent preparation success $p$, observation probability $v$, and symmetric assignment error $\epsilon <1/2$, retain failure and erasure with neutral score $1/2$. The coherent-versus-dephased score contrast is

Uniform complete-law errors $e_0,e_1$ give the retained contrast

The complete-record orientation and sampling conditions are those of Psy02-B CheungPsy02B. A positive contrast certifies sensitivity to the specified phase encoding and detector.

The two phase inputs are orthogonal. A classical memory allowed to store their label can reproduce the binary task. No asymptotic speed, memory, query, or sample advantage is proved. The pilot also does not identify neuronal quantum resources.