Long Live Fine-Tuning: Task-Specific Transformers Outperform Zero-Shot LLMs for Misinformation Response Classification on Reddit

JooYoung Lee, Lin Tian, Angela Brillantes +2

As large language models (LLMs) become default tools for online information verification, an implicit assumption follows them: that scale and general capability are sufficient for…