LLMs used to detect spin in clinical trials, outperforming baselines

Researchers have developed a system using large language models (LLMs) to automatically detect "spin" in clinical trial reporting, specifically focusing on outcome switching. The system, which involves prompt engineering and classification based on token probabilities, achieved an F1 score of 0.78 and 0.90 accuracy on a test set. While outperforming baseline text similarity models, it did not match the performance of fine-tuned BERT models. The LLMs were also used to generate explanations for the detected instances of spin. AI

IMPACT This research demonstrates a novel application of LLMs for ensuring transparency and accuracy in medical research reporting.

RANK_REASON Academic paper detailing a new method for detecting spin in clinical trials using LLMs. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.CL →

• alphaXiv
• arXiv
• BERT
• CatalyzeX
• clinical trials
• Connected Papers
• CORE Recommender
• DagsHub
• Gotit.pub
• Hugging Face
• large-language models
• Litmaps
• ScienceCast
• scite Smart Citations

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →