Using LLMs to identify discourse derailment as a potential cue for disinformation in social media postsOctober 24, 2026