datasets i compile and release, each with full documentation of how it is collected, built, and labeled.
A daily-growing corpus of Weibo hot-search discourse for sociolinguistic research into fandom and online-community language, labeled for validity, target, stance, emotion, moral framing, and satire.
