sejm_committee_transcripts: card says the 454 pre-dedup count included checksum failures

The pipeline's pre-dedup redaction counter also counted PESEL/NIP/REGON
candidates that failed the checksum and stayed in the text: 10 of them in
the 653,411 deduplicated rows of the PR #37 shard. The shipped 403 figure
counts placeholders and is unaffected.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

data/sejm_committee_transcripts/sejm_committee_transcripts.md

@@ -73,7 +73,10 @@ reproducible from the script alone. Requires `pdftotext` (poppler).
   Counted in the shipped shard: **403 redactions** (396 `[ADRES_EMAIL]`, 6
   `[DANE_NUMERY]`, 1 `[DANE_KONTAKTOWE]`; the pipeline's pre-dedup counter read
   454 because it also counted rows later dropped as joint-sitting duplicates or,
-  since the speaker filter, as statements by non-official speakers).
+  since the speaker filter, as statements by non-official speakers, and because
+  it counted PESEL/NIP/REGON candidates that failed the checksum and were left
+  in the text - at least 10 of the 454; the script now counts only actual
+  replacements).
   This is **not**
   [SlayerLab/NERGAL](https://huggingface.co/SlayerLab/NERGAL): the shard must be
   re-run through NERGAL before it is admitted to a release (it quotes citizens'