From f8f778255c4327ac564e11d141439a1d9433e7a3 Mon Sep 17 00:00:00 2001 From: YurenHao0426 Date: Wed, 29 Jul 2026 19:12:11 -0500 Subject: Remove X-not-Y contrast constructions (style) Co-Authored-By: Claude Fable 5 Claude-Session: https://claude.ai/code/session_014FAPDWQ49M5Ye3NpTndTpn --- index.html | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'index.html') diff --git a/index.html b/index.html index 9c2210c..5870fb8 100644 --- a/index.html +++ b/index.html @@ -198,7 +198,7 @@
  • Verified gradient fidelity: the EP update maintains cosine ≈0.99 to the true backprop gradient throughout training: measured, at scale. Larger non-backprop transformers in the literature train most parameters with local backprop inside blocks; ours use none.
    • -
  • Hardware-ready by measurement, not assumption: 8-bit quantization shows no EP-specific +
  • Hardware compatible, with measurements to show it: 8-bit quantization shows no EP-specific penalty against an equally-quantized backprop twin; under injected analog faults (1% forward noise, 10% error-channel noise, device-tolerance mismatch) the EP estimator tracks the faulted network at cosine ≈0.97: learning co-adapts to the hardware.
    • @@ -302,7 +302,7 @@

    Roadmap

    Staged scaling with matched BP controls and hardware-relevant ablations at every rung: a 150M-600M - ladder (does the gap grow or shrink with scale: measured, not assumed), then 1B-3B; each stage + ladder (the gap is measured against a backprop twin at every rung), then 1B-3B; each stage gated on the previous stage’s loss, alignment, and throughput numbers. In parallel: the algorithm→regime map across the activity-difference family (contrastive / coupled-learning arms on the same harness), and a bounded single-column hardware feasibility study.

    -- cgit v1.2.3