New features help detect and guide LLM-generated Korean poetry

Researchers have developed a method to detect and guide the generation of Korean poetry by large language models (LLMs). The approach uses interpretable form-level linguistic features, such as output length, line-final form diversity, line length irregularity, and adherence to standard orthography. This method achieved an 83.60% AUC-ROC for detecting LLM-generated poetry, outperforming existing baselines. Furthermore, expert evaluations showed that LLM-generated poems guided by these features were preferred over unconstrained outputs, with targeted statistics moving closer to human-written poetry distributions. AI

IMPACT This research offers a method to improve the quality and authenticity of LLM-generated creative text, potentially impacting content generation tools.

RANK_REASON Academic paper detailing a new method for detecting and guiding LLM-generated content. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.CL →

• alphaXiv
• arXiv
• CatalyzeX
• DagsHub
• Gemini 3
• Gotit.pub
• GPT-5.2
• Hugging Face
• Influence Flower
• KatFishNet
• Korean poetry
• ScienceCast

• paper
• model release

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