Researchers have explored the capability of large language models (LLMs) to detect narratives in social messages without prior training data. Experiments using human-written narrative descriptions significantly boosted LLM performance on datasets like Dipromats and SemEval, outperforming few-shot or automatically generated descriptions. The study found that ensemble methods, particularly majority voting, improved robustness, and larger models yielded the best results while being less sensitive to prompt variations. These findings suggest LLMs can be a scalable solution for narrative detection, even rivaling supervised systems when human-provided descriptions are used. AI
IMPACT Demonstrates potential for LLMs to scale narrative detection in social media analysis without extensive labeled data.
RANK_REASON Research paper published on arXiv detailing LLM capabilities. [lever_c_demoted from research: ic=1 ai=1.0]
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
