Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

FrameRef: A Framing Dataset and Simulation Testbed for Modeling Bounded Rational Information Health

Victor De Lima, Jiqun Liu, Grace Hui Yang

TL;DR

A simulation-based framework for modeling sequential information exposure and reinforcement dynamics characteristic of ranking and recommendation systems, and construct framing-sensitive agent personas by fine-tuning language models with framing-conditioned loss attenuation, inducing targeted biases while preserving overall task competence.

Abstract

Information ecosystems increasingly shape how people internalize exposure to adverse digital experiences, raising concerns about the long-term consequences for information health. In modern search and recommendation systems, ranking and personalization policies play a central role in shaping such exposure and its long-term effects on users. To study these effects in a controlled setting, we present FrameRef, a large-scale dataset of 1,073,740 systematically reframed claims across five framing dimensions: authoritative, consensus, emotional, prestige, and sensationalist, and propose a simulation-based framework for modeling sequential information exposure and reinforcement dynamics characteristic of ranking and recommendation systems. Within this framework, we construct framing-sensitive agent personas by fine-tuning language models with framing-conditioned loss attenuation, inducing targeted biases while preserving overall task competence. Using Monte Carlo trajectory sampling, we show that small, systematic shifts in acceptance and confidence can compound over time, producing substantial divergence in cumulative information health trajectories. Human evaluation further confirms that FrameRef's generated framings measurably affect human judgment. Together, our dataset and framework provide a foundation for systematic information health research through simulation, complementing and informing responsible human-centered research. We release FrameRef, code, documentation, human evaluation data, and persona adapter models at https://github.com/infosenselab/frameref.

FrameRef: A Framing Dataset and Simulation Testbed for Modeling Bounded Rational Information Health

TL;DR

A simulation-based framework for modeling sequential information exposure and reinforcement dynamics characteristic of ranking and recommendation systems, and construct framing-sensitive agent personas by fine-tuning language models with framing-conditioned loss attenuation, inducing targeted biases while preserving overall task competence.

Abstract

Information ecosystems increasingly shape how people internalize exposure to adverse digital experiences, raising concerns about the long-term consequences for information health. In modern search and recommendation systems, ranking and personalization policies play a central role in shaping such exposure and its long-term effects on users. To study these effects in a controlled setting, we present FrameRef, a large-scale dataset of 1,073,740 systematically reframed claims across five framing dimensions: authoritative, consensus, emotional, prestige, and sensationalist, and propose a simulation-based framework for modeling sequential information exposure and reinforcement dynamics characteristic of ranking and recommendation systems. Within this framework, we construct framing-sensitive agent personas by fine-tuning language models with framing-conditioned loss attenuation, inducing targeted biases while preserving overall task competence. Using Monte Carlo trajectory sampling, we show that small, systematic shifts in acceptance and confidence can compound over time, producing substantial divergence in cumulative information health trajectories. Human evaluation further confirms that FrameRef's generated framings measurably affect human judgment. Together, our dataset and framework provide a foundation for systematic information health research through simulation, complementing and informing responsible human-centered research. We release FrameRef, code, documentation, human evaluation data, and persona adapter models at https://github.com/infosenselab/frameref.
Paper Structure (27 sections, 2 equations, 3 figures, 8 tables)

This paper contains 27 sections, 2 equations, 3 figures, 8 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Dataset Construction
  4. Claim Framing Dimensions
  5. Dataset Workflow
  6. Claim Reframing
  7. Verification
  8. Parsing
  9. Simulation Methodology
  10. Interaction Dynamics
  11. States, Actions, and Policy
  12. Transition Dynamics
  13. Decision Evaluation
  14. Information Health
  15. Agent Personas
  16. ...and 12 more sections

Figures (3)

  • Figure 1: Agent personas assign framing-dependent scores to the same underlying claim.
  • Figure 2: System diagram describing the simulation framework.
  • Figure 3: Authoritative framing tuning results. Left: Task performance as training scale increases ($\alpha=1$). Right: Performance and variability across $\alpha$ values.