Low visibility of Community Notes in Japan during recent election despite rollout of Gaussian scoring model

In early August 2026, the Community Notes system was expanded to use the GaussianModel for Expansion Groups. This was expected to substantially increase the number of helpful notes shown to users globally, including in Japan.

However, during local election in Japan between August and early September 2026, I observed little apparent change in note visibility.

In particular, there were numerous election-related posts containing claims that were subsequently fact-checked as false or misleading by established Japanese news organizations. Some of these posts received very large numbers of views.

For example:
X post https://x.com/zaazasu/status/2093300764328493545?s=20
proposed notes https://x.com/i/birdwatch/t/2093300764328493545?source=6
fact-check https://newsdig.tbs.co.jp/articles/-/2940123?display=1
https://www.okinawatimes.co.jp/articles/-/1916694
https://ryukyushimpo.jp/national/entry-5497523.html

These articles independently fact-checked the claims and identified the central claim as false or misleading.

Nevertheless, despite a substantial number of Community Notes being submitted on election-related posts, I did not personally encounter a single publicly visible Community Note on an election-related post during this period.

This raises the question of whether the GaussianModel expansion is working as intended for Japan, particularly during high-traffic and highly polarized civic events.

Start by tracing the GaussianModel expansion for Expansion Groups and the path that determines whether notes become publicly visible in Japan. Compare the reported election-related submissions and proposed notes with the system's eligibility and visibility results. Done means identifying whether the rollout is functioning as intended or documenting a reproducible failure with supporting evidence.