docs: TODO.sota 20-23 — heb-large arm, plane transfer, release queue, ORT alignment (root cause corrected) (#271)

Co-authored-by: Ronald Tse <ronald.tse@gmail.com>

Author: ronaldtse

TODO.sota/00-README.md

@@ -56,7 +56,11 @@ hf jobs run --flavor a100-large -d \
 | 16 | 16-parking-lot.md | — | — | parked (not now) |
 | 17 | 17-arabic-register-closure.md | P1 | 11 | r8a RUNNING; r8b launching |
 | 18 | 18-parity-runs.md | P1 | 10 | pending |
-| 19 | 19-lexicon-loader.md | P2 | 13 | pending |
+| 19 | 19-lexicon-loader.md | P2 | 13 | DONE (py#33) |
+| 20 | 20-hebrew-large-arm.md | P1 | 04 | RUNNING (run-026, byt5-large/K=4) |
+| 21 | 21-arabic-plane-transfer.md | P2 | 17 | pending verdict |
+| 22 | 22-runtime-releases.md | P1 | 14 | PREPARED — owner-gated versions |
+| 23 | 23-ort-alignment.md | P1 | 18 | golden regen on ORT 1.29 RUNNING; re-dispatch after |
 
 ## Schedule
 

TODO.sota/20-hebrew-large-arm.md

@@ -0,0 +1,10 @@
+# WO20 — Hebrew byt5-large arm (run-026)
+
+RUNNING (a100-large): byt5-large encoder, 3 epochs (large overfits 50K
+units faster), K=4, bs4/accum4. Same gate protocol as WO04.
+
+- Gate: artifact-level DER < 8.18 (beat base) — else negative verdict.
+- Ship: export int8, package, 3-leg parity (goldens on linux/ORT-1.29
+  per WO23), owner-confirmed version.
+- Honest risk: 1.2B params on 50K units may overfit by epoch 2; the
+  per-epoch gate decides.

TODO.sota/21-arabic-plane-transfer.md

@@ -0,0 +1,7 @@
+# WO21 — Arabic plane register transfer (post-r8)
+
+Once WO17's dual-surface verdict names the winning silver mixture
+(r8a self / r8b qcri / r8c both), retrain ara-diac-plane-large with
+that mixture (same plane recipe) so the ON-DEVICE family inherits the
+register closure — not just the dedicated line. Gate: plane-large OOD
+(18.69/11.35) moves AND in-domain stays ≤3.0.

TODO.sota/22-runtime-releases.md

@@ -0,0 +1,11 @@
+# WO22 — runtime release queue (OWNER-GATED, prepared)
+
+Everything staged; version numbers are owner decisions (standing rule):
+
+1. interscript-ts 5.7.0 (plane kind + preserve + parity leg) — unblocks
+   the api worker's real dependency (vitest alias is dev-only today).
+2. interscript-ruby gem next (plane preserve + index-v9 + bound knob).
+3. interscript-py (pip) — still blocked on the previous developer's PyPI
+   trusted-publisher fix; everything staged per TODO.final/01.
+4. api-worker deploy: after (1), drop the vitest alias + type
+   augmentation and deploy with the real package.

TODO.sota/23-ort-alignment.md

@@ -0,0 +1,13 @@
+# WO23 — ORT generation alignment (the parity root cause, corrected)
+
+WO18 mis-attributed the ruby-vs-py 9.25% skew to "the gem's older ORT".
+The gem bundles ORT **1.29.0** — NEWER than the golden-generation
+platform (py 1.20.1). The skew is a GOLDEN-PLATFORM version mismatch.
+
+Fix (eval infrastructure only — no library requirement changes):
+1. Golden generation pins onnxruntime==1.29.0 (HF cpu job).
+2. neural-parity py leg pins the same.
+3. Re-dispatch with DEFAULT bounds; if ruby passes, the calibrated
+   tiers stand and the per-dispatch override reverts to escape-hatch.
+4. ts onnxruntime-node bump (1.27→1.29) is a runtime requirement change
+   — owner decision, only if the ts leg misses tolerance.