Publish the clean-split retrain as revision clean-2026-08-21, and say which revision the numbers describe

README.md

@@ -128,6 +128,31 @@ judgment-domain result, not a "legal English" result, and the difference is exac
 buyer needs to know. For statutory text, use the purpose-built sibling
 [`embed-statute-en`](https://huggingface.co/quanfire-ai/embed-statute-en).
 
+## Which weights should I use?
+
+Two revisions are published. They share the recipe and differ in **what can be said about
+them**:
+
+| Revision | Trained on | Has a valid score? |
+|---|---|---|
+| `main` (default) | 92,419 pairs — the full mined pool | ❌ **No.** It trained on the evaluation documents, so no clean held-out set exists for it *within this corpus* |
+| `clean-2026-08-21` | 46,115 pairs — document-isolated split | ✅ **Yes** — every figure in *Results* above is this adapter |
+
+```bash
+hf download quanfire-ai/embed-legal-en --revision clean-2026-08-21 --local-dir embed-legal-en-clean
+```
+
+**Which one to take.** If you need a number you can cite or audit, take
+`clean-2026-08-21` — it is the one the Results section describes. `main` saw
+2x the training data and may well be the stronger retriever in practice, but "may
+well be" is precisely the kind of claim this card no longer makes.
+
+**We have not swapped the default**, and the reason is worth stating: doing so would trade a
+plausibly-stronger model for a measurable one with no evidence that the trade is good.
+Settling it properly needs a head-to-head of the two adapters on a corpus *neither* of them
+trained on. That is planned, and until it runs, both revisions stay up and this section stays
+honest about which is which.
+
 ## Usage
 
 Pull the adapter and run it through the Quanfire framework, which applies the LoRA