cite the paper-v1.0 tag; correct the audit deviation count

README.md

@@ -274,7 +274,7 @@ Pinned environment — `requirements.txt`, plus:
 | NumPy / scikit-learn / tokenizers / safetensors | 2.4.2 / 1.8.0 / 0.22.2 / 0.7.0 |
 | GPU / driver | NVIDIA H200, driver 550.90.07 |
 | tokenizer SHA-256 | `8bbb4fd7db2928b323b7cb7ba595e88130f49a717fb7a79b3bd33b4b5e248fcf` |
-| dataset revision | `nur-dev/pns-world` @ `PLACEHOLDER_DATASET_REVISION` |
+| dataset revision | `nur-dev/pns-world` @ `paper-v1.0` |
 | seeds | 1, 2, 3 (recorded per checkpoint in `config.json`) |
 
 Config and artifact hashes: `results/artifact_hashes.csv`.
@@ -290,8 +290,7 @@ narrative state works when scaled. The Experiment 2 control is capacity- and
 parameter-matched to PNS-Bind but is **not** architecture-matched to
 Experiment 1's PNSR — which of those differences makes an unstructured state
 functional at all is not isolated by any experiment here. `AUDIT.md` in the
-dataset repo records nine deviations, two of which materially changed
-interpretation.
+dataset repo records eleven deviations, four of which bear on a claim.
 
 ## License
 
@@ -305,6 +304,6 @@ See `CITATION.cff`.
 @article{pnsbind2026,
   title  = {Beyond Transcript Replay: Causal Persistent Memory Through Protected Semantic Bindings},
   year   = {2026},
-  note   = {Code and models: https://huggingface.co/nur-dev/pns-bind-25m (revision PLACEHOLDER_MODEL_REVISION)}
+  note   = {Code and models: https://huggingface.co/nur-dev/pns-bind-25m, tag paper-v1.0}
 }
 ```