ResearchBenchmarksDisinformation Detection on post-cutoff datasetsDisinformation Detection on post-cutoff datasetsFollow87.7F1 ScorePCoT46.93257.51668.178.684Jun 7, 2025Evaluation resultsMethodMethodLinksF1 ScorePCoTmodel=Gemini 1.5 Flash...model=Gemini 1.5 Flash, runs_averaged=32025.0687.7PCoTmodel=GPT 4o mini, run...model=GPT 4o mini, runs_averaged=32025.0687.3PCoTmodel=Llama 3.3 70B, r...model=Llama 3.3 70B, runs_averaged=32025.0686.4PCoTmodel=Llama 3.1 8B, ru...model=Llama 3.1 8B, runs_averaged=32025.0679.4PCoTmodel=Claude 3 Haiku,...model=Claude 3 Haiku, runs_averaged=32025.0678.3BERTmodel=BERTmodel=BERT2025.0648.5