Interactive simulation of a governed inference pipeline
Press T to follow one token through the reactor, 1 to 8 to inject a fault, or select any ring or invariant to see what it is, the real number or "simulated", and its source.
From the MIND inference repository, gated by MIND-Law, STARGA's implementation of the 512 constraint by Dr. J.M. Watson
A governed inference pipeline, drawn ring by ring. Every number is measured and cited, or labelled simulated.
A simulation, not a running model. No weights are loaded and the output text is a fixed script. The kernel predicates are re-implemented here in JavaScript, two published MIND-Law digests are shown as constants, and every record hash is a real SHA-256 over a simulated record. The figures in the Numbers tab are quoted from the MIND inference repository's benchmark results, with their caveats.