Replace coming-soon pcloss stub with live HF link

README.md

@@ -58,8 +58,10 @@ artifacts:
 
 - **Code:** [`mirrorethic/t3-reference`](https://github.com/mirrorethic/t3-reference) (Apache-2.0)
 - **Trace library + benchmarks:** <https://t3atlas.dev>
-- **Sibling checkpoint:** `mirrorethic/t3-124m-v36-pcloss` *(coming soon)* — same
-  weights but trained with the inter-stage predictive-coding loss un-detached.
+- **Sibling checkpoint:** [`mirrorethic/t3-124m-v36-pcloss`](https://huggingface.co/mirrorethic/t3-124m-v36-pcloss) — same
+  architecture, trained with the inter-stage predictive-coding loss un-detached.
+  Slightly worse PPL (28.53 vs 27.76), neutral on reasoning. Use the pair for
+  the controlled inter-stage-PC ablation.
 
 ## Quick start
 
@@ -171,11 +173,11 @@ compute per token (S1 averages 2.7–3.7 ponder loops on these tasks).
 - Trained without grade-3 trivector terms (the static bivector Ω is the
   full Cl(3,3) rotation in this checkpoint). The state-dependent variant
   is a planned v3.7 Medium-scale follow-up.
-- The `v36_pcloss` sibling at <https://huggingface.co/mirrorethic/t3-124m-v36-pcloss>
-  *(when uploaded)* is the same architecture trained with the inter-stage
-  predictive-coding loss un-detached. Slightly worse PPL (28.53), neutral
-  on reasoning — the K-predictor learning a real cross-stage map (r=0.59)
-  doesn't translate into downstream gains at this scale.
+- The [`v36-pcloss` sibling](https://huggingface.co/mirrorethic/t3-124m-v36-pcloss)
+  is the same architecture trained with the inter-stage predictive-coding
+  loss un-detached. Slightly worse PPL (28.53), neutral on reasoning — the
+  K-predictor learning a real cross-stage map (r=0.59) doesn't translate
+  into downstream gains at this scale.
 
 ## Capabilities probe