diff options
Diffstat (limited to 'NOTE.md')
| -rw-r--r-- | NOTE.md | 24 |
1 files changed, 23 insertions, 1 deletions
@@ -5,7 +5,7 @@ - **pilot**: Controlled iteration (commits 0b9ebb2, 7baf7ae) - **frozen**: Code at commit 0b9ebb2 for all reported results -## Status: PHASE 9 COMPLETE — OFFLINE PREFIT + BLEND IS THE KEY +## Status: PHASE 10A — NO PREFIT THRESHOLD, BLEND ITSELF IS THE ACTIVE INGREDIENT --- @@ -550,3 +550,25 @@ The +1.5% gain from 9A's blend(0.75) at t0=5 is the project's best online result - `checkpointed_handoff/`: Phase 9A checkpointed handoff with branches - `periodic_refit/`: Phase 9B periodic refit - `topdown_curriculum/`: Phase 9C top-down curriculum + +--- + +## Phase 10A: Prefit Threshold Curve + +**Setup**: t0=5, blend_075, E_prefit in {0, 15, 60}, seed=42 + +| E_prefit | Gamma_frozen | rho_frozen | final acc | diff vs DFA | +|----------|-------------|-----------|-----------|-------------| +| 0 (random Vec) | -0.005 | 0.014 | **0.324** | **+1.3%** | +| 15 | 0.002 | 0.011 | 0.323 | +1.2% | +| 60 | -0.001 | -0.009 | 0.325 | +1.4% | +| continue_DFA | — | — | 0.311 | baseline | + +**Case C: NO prefit threshold exists.** Even random Vec (E=0) with blend(0.75) beats DFA. + +**Critical reinterpretation of Phase 9A**: the +1.5% gain was NOT from Vec learning good credit. +Frozen Gamma/rho are near zero at all prefit levels. The benefit comes from the blend mechanism +itself — blending DFA with any additional signal provides regularization/diversification. + +### Experiment IDs (Phase 10) +- `prefit_threshold/`: Phase 10A prefit threshold curve |