Experiments

V30: Dual Prediction

Negative result. Dual prediction (self + social) is worse than social-only.

Substrate: MLP with 2 outputs. Loss = MSE_self + MSE_social. Mean $\intinfo = 0.091$ (vs 's 0.104).

Self MSE is 100-150x smaller than social MSE. The easy self target colonizes the shared representation, destroying 's richness path. Lesson: naive multi-task learning fails when targets have vastly different difficulty scales.

Source code

Study record — canonical metadata, result path, status, seeds, and key finding.

v30_substrate.py — Dual-output MLP
v30_evolution.py — Evolution loop
v30_gpu_run.py — GPU runner