A Multi-Stage Agentic Framework for Effective Counter-Narrative Generation and Refinement
A multi-agent refinement process that iteratively improves LLM-driven counter-narratives for persuasiveness, emotional engagement, and shareability is introduced, highlighting a pathway toward scalable, narrative-specific interventions against hate speech and misinformation.
Carmel Kronfeld, Sharva Gogawale, Tetsuro Kobayashi et al.
· 0 citations