Update persona template library dataset

clarify off-axis score and template provenance

expand confound audit docs

add source urls and confound list

use normalized score components

make main table human-facing

simplify intro and public splits

simplify public dataset splits

add clean score tables

simplify card and remove evidence split

switch dataset viewer to parquet splits

Add measured v2 pilot stats

Add v2 candidate persona library

Upload v1 persona steering template library

initial commit