Study: LLM text may increase cognitive effort per word despite shorter length

A new study published on arXiv introduces RoBERTa-RT, a model designed to predict reading times, to analyze the cognitive load of reading text generated by Large Language Models (LLMs) compared to human-written text. The research found that while LLM-generated rewrites are often shorter and have lower overall predicted reading times, they exhibit higher predicted reading times per word. These differences are linked to linguistic variations such as longer words, increased lexical diversity, and higher contextual predictability in LLM outputs. The findings suggest that LLM writing may shift cognitive demands, reducing the overall text volume but increasing the processing effort required for individual words. AI

IMPACT LLM-generated content may require more cognitive effort per word, potentially impacting reader comprehension and engagement.

RANK_REASON The cluster contains a research paper published on arXiv detailing a new model and its findings on LLM text. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.CL →

• alphaXiv
• arXiv
• CatalyzeX
• DagsHub
• Gotit.pub
• Hugging Face
• human
• Large Language Models
• RoBERTa-RT
• ScienceCast

• paper
• other

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