A dashboard is a set of dials that help a pilot tell what a system is doing. This report builds such a dashboard for a real-time cognitive runtime, using measurements that are fast to compute from logs and hard to game by trivial scaling. The goal is not to argue that any single number is consciousness; the goal is to provide an instrument that makes claims about “more structured” versus “more random” versus “more coupled” states testable. Technically, a 15040-tick trajectory is analyzed (two regimes: #0 for 579 ticks and #4 for 14461 ticks). A sliding-window dashboard combines (i) permutation entropy Bandt and Pompe, 2002, (ii) Lempel–Ziv (LZ) complexity Lempel and Ziv, 1976, Kaspar and Schuster, 1987, (iii) spectral scaling (Welch PSD slopes), (iv) fractal/embedding proxies (DFA and correlation dimension), and (v) a directed-influence proxy based on lag-5 Granger tests Granger, 1969, Barnett and Seth, 2014. The instrument cleanly distinguishes the two regimes in this run: the Granger-causal density rises from 0.0833 to 0.75, and the z-scored multi-metric heatmap shows a coordinated shift in visit-structure and complexity indicators. A lightweight perturbational complexity proxy (EPCI) is also defined from endogenous “event” windows, as a scalable analogue of perturbational complexity indices Casali et al., 2013. All quantitative statements are bounded to the included artifacts and should be treated as a T2 instrument result with a T3 smoke demonstration (single-run; not preregistered generalization).
Justin Lietz (Sat,) studied this question.