mdok-style at SemEval-2026 Task 10: Finetuning LLMs for Conspiracy Detection

arXiv:2605.02712v1 Announce Type: cross Abstract: SemEval-2026 Task 10 is focused on conspiracy detection. Specifically, the goal is to detect whether a Reddit comment expresses a conspiracy belief. Our submitted mdok-style system utilizes data augmentation and self-training (to cope with a rather...