Glean Health — Corpus Overview

Aggregate view across every condition. Source: data/pipeline/*_pipeline_output.json. Generated by scripts/build_corpus.py.

Per-condition breakdown

Sortable. Click a column header to sort. Click a condition name to open its detailed dashboard.
Condition Extractions Mentions Clusters Top novelty surfaces

Author cross-condition distribution

How many authors post in N conditions (out of 25). Almost all authors post in only 1.

Cross-condition authors (2+)

Sortable. Click a row to expand and see what they posted in each condition. Confirmed-shill rows are highlighted.
Author Conditions Total mentions Conditions list Detail

Cluster "poisoning" view

Per-condition, the most concentrated clusters by three measures.
Top-author share (orange): max single-author mentions / cluster mentions. High = one person dominates the cluster.
Mega-poster share (purple): mentions from authors with ≥10 corpus-wide mentions / cluster mentions. High = cluster is amplified by power-users.
Cross-condition share (red): mentions from authors posting in 3+ conditions / cluster mentions. High = wellness-cluster gravity is pulling.