From 57ad5f340a5148417d5f50420af0f7982ddc8105 Mon Sep 17 00:00:00 2001 From: Yuren Hao Date: Thu, 9 Jul 2026 04:31:59 -0500 Subject: cascade: zil interleaved scheme = exact gradients at all depths (cos 1.0000 L6-24, trajectory 0.9998+); casc_ep_train zil trainer; naive-relaxation depth failure documented Co-Authored-By: Claude Fable 5 Claude-Session: https://claude.ai/code/session_014FAPDWQ49M5Ye3NpTndTpn --- docs/campaign/CASCADE_ABLATION_PLAN.md | 20 ++++++++++++++++++++ 1 file changed, 20 insertions(+) (limited to 'docs') diff --git a/docs/campaign/CASCADE_ABLATION_PLAN.md b/docs/campaign/CASCADE_ABLATION_PLAN.md index 5c10dd9..c480c11 100644 --- a/docs/campaign/CASCADE_ABLATION_PLAN.md +++ b/docs/campaign/CASCADE_ABLATION_PLAN.md @@ -26,6 +26,26 @@ The looped-EP precedent multiplier was ~230× BP; the K-frontier decides whether --- +## STATUS 2026-07-09 (same day): Tier 0 CLOSED GREEN via the zil scheme; C1 running + +- **Naive relaxation FAILS at depth** (the B1-lite sweep): jacobi K=40·L → cos 0.82 (L6) / 0.67 (L12) + / 0.53 (L24), shrink dying 0.41→0.28; gsf/gsr with small-η+momentum no better; β-insensitive + (0.01/0.03/0.1 identical) ⟹ binding error = RELAXATION INCOMPLETENESS, not Taylor bias. + warp2.0 catastrophic (cos 0.11) under naive descent. +- **Two implementation traps found:** (1) NBT-normalized energy made γ=1 actually γ=1/128; + (2) plain γ=1 reverse sweep WITHOUT interleaved reads contaminates e_l with J_l·δ_{l−1} + (same β-order as the signal) — final-state readout is directionally ruined (cos 0.30@L6). +- **The fix = zil scheme (interleaved reverse sweep):** update z_l (γ=1, SUM units) then read + θ_l IMMEDIATELY (e_l = −β·δ_l exact at the feedforward point; δ-recursion has NO linearization + error). Single phase, β cancels exactly. **Results: cos = 1.0000 at L=6/12/24; io gate 0.9999; + warp2.0 → 1.0000; real-trajectory ckpts (casc_bp6 s0→s4000) → 0.9998–1.0000. A0.1/A0.2/A0.3 all + green.** Honest framing: zil is numerically BP restructured as per-layer local two-factor energy + reads (no global backward graph); the EQUILIBRIUM mode (jacobi/CG to convergence) remains the + physically-meaningful EP column — priced expensive by the sweep, CG/preconditioning is the B2 job, + and it is the analog-hardware rung (E-tier). +- **C1 money run LAUNCHED:** casc_ep6 (zil, seed 0) vs casc_bp6 (BP twin, best 2.9746@4k) — + step-0 CE bit-identical (true twin). 107 GPU3. + ## Tier 0 — gate hardening (probe-scale, hours, no training) → K1 | ID | question | design | decision rule | -- cgit v1.2.3