ContentFuzz rewrites posts with LLM guidance from stance model confidence to flip machine labels without altering human intent, tested across four models and three datasets in two languages.
Bowen Zhang, Daijun Ding, Liwen Jing, Genan Dai, and Nan Yin
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
fields
cs.CL 2years
2026 2verdicts
UNVERDICTED 2representative citing papers
Prompt-based methods outperform agent-based debate for LLM stance detection, with model scale driving larger gains than method choice and reasoning models showing no consistent edge.
citing papers explorer
-
Content Fuzzing for Escaping Information Cocoons on Digital Social Media
ContentFuzz rewrites posts with LLM guidance from stance model confidence to flip machine labels without altering human intent, tested across four models and three datasets in two languages.
-
A Systematic Comparison of Prompting and Multi-Agent Methods for LLM-based Stance Detection
Prompt-based methods outperform agent-based debate for LLM stance detection, with model scale driving larger gains than method choice and reasoning models showing no consistent edge.