Characterize the response function over continuous HGR levels

Characterize the response function relating realized Historical Grounding Ratio levels to experiential outcomes across a continuously sampled HGR range, and determine whether the apparent advantage of an intermediate grounding level is stable and whether different experiential goals correspond to distinct appropriate HGR ranges.

Background

The experiment compared only situation-dominant, balanced, and evidence-dominant conditions, which formed discrete low, medium, and high regions rather than a continuous range of realized HGR values. Achieved HGR also varied within conditions, so prompt targets cannot be equated directly with output values. The paper therefore leaves unresolved the shape of the experience–HGR relationship and whether the intermediate condition's apparent benefit generalizes.

References

The experiment used three discrete conditions to form low, medium, and high regions of historical grounding. This enabled clear within-subject comparisons but cannot identify the response function across a continuous HGR space. Achieved HGR also varied within conditions, so target configurations in prompts cannot be equated directly with realized values. Future work could specify more target grounding levels, sample HGR more densely with larger samples, and verify the manipulation using measured values from delivered texts. Such studies could test whether the apparent advantage of an intermediate level is stable and whether different experiential goals correspond to different appropriate ranges.

Balancing Evidence and Interpretation: Historical Grounding Ratio as a Design Parameter for AI-Generated Urban Storytelling  (2608.24157 - Zhang et al., 25 Aug 2026) in Research Boundaries and Future Directions, Section 6.2

The study therefore examines system-directed source allocation at generation time rather than a full conversational process in which users and the system negotiate narrative depth. Participants' expectations of additional interaction methods suggest that the effect of preset HGR may change once users can actively request more historical material, ask for explanations of visible objects, or redirect the narrative. Future work could extend HGR from a parameter for a single generation into a dynamic state jointly regulated by user behavior and context, enabling study of how system control and active user control together determine source configuration.

Balancing Evidence and Interpretation: Historical Grounding Ratio as a Design Parameter for AI-Generated Urban Storytelling  (2608.24157 - Zhang et al., 25 Aug 2026) in Research Boundaries and Future Directions, Section 6.6