Experiments
Gutenberg Curriculum Baseline
draft
Training on curriculum-aligned Gutenberg narrative text will reduce loss steadily without early collapse into repetition, and will improve short prompt continuation coherence relative to a minimal smoke dataset.
Curriculum Ordering Baseline (Dataset Plan Locked)
draft
A curriculum-ordered training schedule will alter training stability and/or downstream behavior relative to a shuffled control, even when total token count and model architecture are held constant.
EXP-0004 — Guided Choice / Preference Signal (Adaptive Sampling)
draft
If we allow the organism to bias its dataset sampling using a simple adaptive rule, then training and behavior will diverge measurably from fixed mixing weights, because the organism’s learning dynamics will express a stable preference sig…
EXP-0003 — “Play” via Controlled Novelty (Small Perturbations)
draft
If we introduce a small, controlled novelty pressure (e.g., lightly perturbed text during sampling, or prompt mutation during evaluation), then the organism will become more robust to minor corruption without destabilizing training, becaus…
EXP-0002 — Complexity Beats Volume (D0 vs D0+D1)
draft
If we inject slightly more complex narrative structure (D1) into the baseline (D0) via a controlled mix, then late-stage learning will improve faster than simply training longer on D0 alone, because broader structure reduces saturation and…
EXP-0001 — Repeatability Band (Seed Variance)
draft
If we train on the same baseline dataset (D0) with only the random seed changed, then the organism’s learning curves and early behavior snapshots will fall within a narrow variance band, because the training loop and data distribution are…