"Scaling Prediction for Global-Workspace Emergence in Language Models"
Open MIND July 15, 2026 Jeremy Clements
A preregistered prediction tests whether workspace-like internal representations in language models arise from a fundamental consciousness field (the Prima Conscientia framework) or from complexity-driven phase transitions. Under the field view, proto-workspace structure should appear even in very small models and strengthen smoothly with scale; under the emergence view, workspace properties should appear abruptly at capability thresholds. The test uses the Jacobian-lens methodology and the Pythia model suite, with resolution controls and falsification conditions. The prediction is placed on the public record before measurements exist.