analiza aiqbench
This research establishes a rigorous evaluation framework for auditing LLM-based stance simulation by testing sensitivity to counterfactual context revisions. The findings demonstrate that simulated user opinions undergo robust and effective shifts even with text-only or multimodal context changes, independent of specific polarization mechanisms. For practitioners, this underscores the critical need to account for context fragility when deploying models to predict or influence online opinion dynamics.
Revising Context, Shifting Simulated Stance: Auditing LLM-Based Stance Simulation in Online Discussions
arXiv:2606.06443v1 Announce Type: new Abstract: Large language models are increasingly used to simulate social media users and infer how individuals may respond to online discussions. However, it remains unclear whether these simulations reflect precise user-specific beliefs or whether they are highly sensitive to semantically independent changes in conversational contexts. In this work, we study counterfactual context revision as a framework for auditing LLM-based stance simulation. Given an original online conversation, we first infer a target user's stance toward a specific topic. We then apply controlled revision strategies to the conversational context and simulate the user's stance again under the revised context. We compare text-only revision strategies with a multimodal one that incorporates meme-based context and evaluate two main effectiveness metrics, i.e., average directional stance shift and stance transition rate. The results reveal effective and robust stance transitions in both text-only and multimodal strategies across different polarization-preference mechanisms. Our study contributes an evaluation framework for understanding the context sensitivity of LLM-based stance simulation. More broadly, it highlights both the promise and risk of using LLMs to simulate online opinion dynamics.
Abstract:Large language models are increasingly used to simulate social media users and infer how individuals may respond to online discussions. However, it remains unclear whether these simulations reflect precise user-specific beliefs or whether they are highly sensitive to semantically independent changes in conversational contexts. In this work, we study counterfactual context revision as a framework for auditing LLM-based stance simulation. Given an original online conversation, we first infer a target user's stance toward a specific topic. We then apply controlled revision strategies to the conversational context and simulate the user's stance again under the revised context. We compare text-only revision strategies with a multimodal one that incorporates meme-based context and evaluate two main effectiveness metrics, i.e., average directional stance shift and stance transition rate. The results reveal effective and robust stance transitions in both text-only and multimodal strategies across different polarization-preference mechanisms. Our study contributes an evaluation framework for understanding the context sensitivity of LLM-based stance simulation. More broadly, it highlights both the promise and risk of using LLMs to simulate online opinion dynamics.
Computation and Language (cs.CL); Multimedia (cs.MM); Social and Information Networks (cs.SI)
arXiv:2606.06443 [cs.CL]
(or arXiv:2606.06443v1 [cs.CL] for this version)
https://doi.org/10.48550/arXiv.2606.06443 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 276 cuvinte
Subiecte conexe
Revising Context, Shifting Simulated Stance: Auditing LLM-Based Stance Simulation in Online Discussions
arXiv · 9 iun. 2026
Revising Context, Shifting Simulated Stance: Auditing LLM-Based Stance Simulation in Online Discussions
arXiv · 1 sept. 2026
How might continual learning affect safety and alignment?
LessWrong · 13 iun. 2026
[AINews] High Taste Testers Ep. 1
Latent Space · 17 oct. 2026