faeval.git/experiments/minimal_scaffold_replication.py, branch master Unnamed repository; edit this file 'description' to name the repository. Add Phase 10A.8C: 3-seed replication — scaffold gains are marginal 2026-03-27T23:07:58+00:00 YurenHao0426 Blackhao0426@gmail.com 2026-03-27T23:07:58+00:00 2a230acd5ee3fa6605892d524badf281ba7e9cfd 3-seed results (mean±std): - DFA: 0.306±0.006 - perlayer_vector α=0.75: 0.304±0.006 (-0.2%, not significant) - random_trainable α=0.75: 0.313±0.007 (+0.7%, marginal, error bars overlap) Single-seed gains (+1.1% perlayer, +0.8% vec) do not robustly replicate. The scaffold mechanism provides at best a marginal, statistically uncertain benefit. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> 
3-seed results (mean±std):
- DFA: 0.306±0.006
- perlayer_vector α=0.75: 0.304±0.006 (-0.2%, not significant)
- random_trainable α=0.75: 0.313±0.007 (+0.7%, marginal, error bars overlap)

Single-seed gains (+1.1% perlayer, +0.8% vec) do not robustly replicate.
The scaffold mechanism provides at best a marginal, statistically uncertain benefit.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>