Correct: one retention filter, majority label must match the connector

README.md

@@ -100,22 +100,21 @@ The written guidelines required annotators to judge **strictly what is written**
 repairing or extending the sentences, and to choose `neutral` when a pair was ambiguous. No
 synthetic augmentation was used at any stage.
 
-### Two filters, not one
-
-Two selection steps were applied, and both matter when interpreting scores on this set:
-
-1. **Annotator agreement** — pairs without a sufficient majority among the 3 annotators were
-   discarded (151 removed).
-2. **Agreement with the connector label** — only pairs whose human majority label *also* matched the
-   connector-derived label were retained. This is why `connection_type` and `connector_type` are
-   identical on all 1,695 rows.
-
-The second filter means this is a **reliability-filtered slice**, not a neutral sample: it holds
-only pairs where automatic and human labelling already coincide. Scores here are correspondingly
-higher than on the full ESNLIR test split (0.82 vs 0.68 macro F1 for the same XLM-RoBERTa
-checkpoint), and that gap measures **label reliability rather than task difficulty** — the same
-effect documented for ESNLIR's own human-validated subset. Do not read it as models being better at
-these instances.
+### The retention filter
+
+A pair was kept only if its **human majority label matched the connector-derived label**. The 151
+discarded instances are exactly those where the two disagreed — which is why `connection_type` and
+`connector_type` are identical on all 1,695 remaining rows.
+
+This makes the set a **reliability-filtered slice**, not a neutral sample: it holds only pairs where
+automatic and human labelling already coincide. Scores here are correspondingly higher than on the
+full ESNLIR test split (0.82 vs 0.68 macro F1 for the same XLM-RoBERTa checkpoint), and that gap
+measures **label reliability rather than task difficulty** — the same effect documented for ESNLIR's
+own human-validated subset. Do not read it as models being better at these instances.
+
+The retention rate is notably high: 1,695 of 1,846 candidates, 91.8%. ESNLIR's own validation
+exercise retained only ~45% under the same rule, so these candidates — drawn by confidence
+stratification — were substantially easier to label consistently.
 
 ### How candidates were chosen