DeepSeek-R1 Prompt to Audit Causal Claims in Papers
Use DeepSeek-R1 at temperature=0.5 with a single-instruction zero-shot prompt to evaluate statistical validity in draft manuscripts.
API Config: temperature=0.5, top_p=0.95 User Prompt: Evaluate the experimental design and statistical power calculations in the following research draft: {{DRAFT}}. Identify any unstated confounding variables or multiple-hypothesis testing errors.Fill in the blanks (1)
Nobody has confirmed this prompt yet. If you try it, tell others whether it worked.
No comments yet.