arXiv subjects

Heechan Lee

Publications and source records attributed to Heechan Lee.

3 recordsLinked to original sources

arXiv2026

TIDES: A Longitudinal Bilingual Dataset for Modeling Multi-Party Social Dynamics

Group conversations are fundamental to human collaboration, yet standard large language models (LLMs) still struggle with the complexities of multi-party interaction. This challenge persists in part because existing group conversation datasets are often limited to short-term lab settings with contrived tasks, failing to capture the long-term social dynamics of real-world teams. To bridge this gap, we introduce TIDES, a high-resolution longitudinal dataset tracking 12 university project teams over a full semester. Comprising 75,971 utterances in both English and Korean from in-person meetings, TIDES provides a naturalistic record of teams working on self-managed projects. Our socio-structural annotations-covering interaction types, emergent roles, and development stages-allow for modeling of team evolution over months. Experiments show that fine-tuning on TIDES improves next-speaker prediction by 13.8 percentage points over a bigram baseline (64.53%) and yields performance comparable to strong proprietary zero-shot models. The model also comes within 2.1 percentage points of the published state of the art on the AMI Meeting Corpus while using approximately 42% less training data. However, human evaluations suggest that better next-speaker prediction does not necessarily yield more natural or coherent utterances, as fine-tuned models were generally less preferred than vanilla models. This potential mismatch motivates further study of how structural modeling can support natural multi-party generation.

cs.CL↗

arXiv2025

Evalet: Evaluating Large Language Models through Functional Fragmentation

Practitioners increasingly rely on Large Language Models (LLMs) to evaluate generative AI outputs through "LLM-as-a-Judge" approaches. However, these methods produce holistic scores that obscure which specific elements influenced the assessments. We propose functional fragmentation, a method that dissects each output into key fragments and interprets the rhetoric functions that each fragment serves relative to evaluation criteria -- surfacing the elements of interest and revealing how they fulfill or hinder user goals. We instantiate this approach in Evalet, an interactive system that visualizes fragment-level functions across many outputs to support inspection, rating, and comparison of evaluations. A user study (N=10) found that, while practitioners struggled to validate holistic scores, our approach helped them identify 48% more evaluation misalignments. This helped them calibrate trust in LLM evaluations and rely on them to find more actionable issues in model outputs. Our work shifts LLM evaluation from quantitative scores toward qualitative, fine-grained analysis of model behavior.

cs.HC↗

arXiv2025

Radon Exposure Dataset

Exposure to elevated radon levels in the home is one of the leading causes of lung cancer in the world. The following study describes the creation of a comprehensive, state-level dataset designed to enable the modeling and prediction of household radon concentrations at Zip Code Tabulation Area (ZCTA) and sub-kilometer scales. Details include the data collection and processing involved in compiling physical and demographic factors for Pennsylvania and Utah. Attempting to mitigate this risk requires identifying the underlying geological causes and the populations that might be at risk. This work focuses on identifying at-risk populations throughout Pennsylvania and Utah, where radon levels are some of the highest in the country. The resulting dataset harmonizes geological and demographic factors from various sources and spatial resolutions, including temperature, geochemistry, and soil characteristics. Demographic variables such as the household heating fuel used, the age of building, and the housing type provide further insight into which populations could be most susceptible in areas with potentially high radon levels. This dataset also serves as a foundational resource for two other studies conducted by the authors. The resolution of the data provides a novel approach to predicting potential radon exposure, and the data processing conducted for these states can be scaled up to larger spatial resolutions (e.g., the Contiguous United States [CONUS]) and allow for a broad reclassification of radon exposure potential in the United States.

cs.CE↗