X open-sources "recommendation algorithm" code and launches transparency tool, users can view account and post downranking labels

The X platform announced further enhancements to algorithm transparency, officially open-sourcing the code relevant to the exposure ranking of posts in the "For You" recommended timeline, and launching new transparency tools that enable users to check whether their accounts or posts have been assigned labels that may limit visibility. X stated that this move aims to enable users to determine for themselves whether the platform restricts their content distribution, whether the recommendation system is fair, and why they see specific content. The two features released this time were described by X as "complementary pieces of a puzzle": the open-source code demonstrates the generation mechanism of the "For You" timeline, including content tags affecting post exposure and their functions; the transparency tool allows users to check whether their own accounts or posts have restriction labels affecting visibility. It is reported that this feature is still in the pilot phase and is available only to randomly selected eligible accounts. Initial test subjects include users who have been registered for over a year and have posted more than 10 times in the past month. X plans to gradually expand coverage based on user feedback. Third-party recommendation system experts have been invited to review and test this code release in advance to ensure information disclosure is clearer and more practical. The company stated that going forward, as the recommendation algorithm continues to evolve, it will regularly update the relevant transparency mechanisms.