Researchers have developed a Structured Chain-of-Thought (SCoT) prompting method using the open-source Llama3-8B model to assist in the frame analysis of migration news. This approach aims to provide transparent and auditable results for media scholars, addressing concerns about proprietary APIs and equitable access. Evaluations show SCoT improves classification performance and that human annotators found the generated reasoning logical, though it also influenced their judgment. AI
IMPACT Provides a more accessible and auditable method for researchers to analyze media narratives, potentially influencing how LLMs are used in social science.
RANK_REASON The cluster contains an academic paper detailing a new method for LLM-assisted frame analysis.
AI-generated summary · Google Gemini · from 2 sources. How we write summaries →
