Answers are grounded in this paper and cite their sources.
Textual sensemaking in spatial layouts allows users to externalize complex mental models, yet synthesizing these unstructured arrangements into linear narratives remains a severe cognitive burden. Current LLM-assisted tools largely ignore the rich structural and semantic context embedded within 2D space. To address this, we introduce the Space-to-Report framework, a novel paradigm leveraging Space-Augmented Generation (SAG) to translate 2D spatial layouts into steerable constraints for continuous narrative generation. We instantiate this framework in ReSPIRE, a prototype that maps interactive spatial layouts to narrative reports to stimulate human reflection-in-action for iterative sensemaking. An empirical evaluation confirms that SAG achieves predictable, linear controllability over space-report alignment under controlled spatial inputs. Furthermore, a user study ( N = 12) shows that ReSPIRE effectively bridges the gap between human spatial thinking and LLM linear generation, significantly reducing cognitive load while accelerating iteration. Crucially, validated by empirical behavioral logs, the framework drives a fundamental workflow shift toward fluid, iterative sensemaking, positioning the human as a strategic “architect” and the LLM as an aligned “executor” within a shared cognitive space.
No takes yet. Share an insight, caveat, or question.
Tang et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: