arXiv subjects

Junyeong Park

Publications and source records attributed to Junyeong Park.

At least 19 recordsLinked to original sources

arXiv2026

A Representation-Learning Item Response Model for Identifying Behaviorally Important Actions in PIAAC Process Data

Problem-solving log process data from computer-based assessments provide detailed information about how respondents approach and complete tasks. However, the resulting action sequences are complex and noisy, making it difficult to identify specific behaviors associated with successful performance. This paper proposes a representation-learning item response modeling (IRT) framework for identifying behaviorally important actions while accounting for respondent proficiency and item-level differences. Raw log sequences and timing information are first transformed into action representations that incorporate the hierarchical structure of action labels and the sequential and temporal context in which each action occurs. These respondent-specific representations are then entered as covariates in an extended IRT model, with spike-and-slab priors used to identify action-item combinations associated with response accuracy. The framework therefore evaluates actions contextually rather than as simple occurrence indicators and provides posterior uncertainty for their associations with performance. We apply the approach to problem-solving process data from the OECD Programme for the International Assessment of Adult Competencies (PIAAC). The analysis identifies a sparse set of actions associated with successful and unsuccessful performance and reveals differences across items in where behavioral information occurs within the problem-solving process.

stat.AP↗

arXiv2026

Investigating Social Bias in Narrative Image Generation

Text-to-image (T2I) generation models are increasingly embedded in applications such as media content creation and education, raising concerns about how their outputs may reproduce social biases. Prior work has shown that T2I models exhibit social biases, yet existing evaluations largely focus on a photo generation task. As a result, it remains unclear whether and how such biases manifest in more narrative visual formats, such as storyboards and comics, where characters and events are presented across multiple panels. In this work, we compare bias expression across photo, storyboard, and comic generation in six T2I models by adapting BBG, a text-based bias evaluation framework, to image generation. Our results show that proprietary models generate 25.9% biased outputs in photo generation on average, with biased outputs increasing by 9.6pp in storyboard generation and 18.2pp in comic generation. We also find that photos mainly encode biases through subtle visual cues, while storyboards and comics reveal them more explicitly through event sequencing, character positioning, narrative resolution, and textual elements. These findings show that biases that remain less visible in photo generation may surface in narrative visual formats, highlighting the importance of evaluating T2I systems with diverse visual formats beyond photo generation.

cs.CV↗

arXiv2026

EduZone: A Framework for Evaluating LLM Safety for K-12 Students and Teachers

Large language models (LLMs) are increasingly used across diverse tasks in K-12 education, yet existing safety evaluations rarely examine how harmful or inappropriate content appears in interactions between LLMs and students or teachers. To address this, we present EduZone, an evaluation framework for LLM safety across diverse educational scenarios. Our framework systematically combines (1) student- and teacher-facing LLM usage contexts, (2) fine-grained curriculum concepts, and (3) 6 risk categories and 28 subcategories spanning both conventional and education-specific harms to generate contextually grounded adversarial interactions. We construct these interactions in three settings: single-turn requests, static multi-turn conversations, and dynamic multi-turn conversations. Using these interactions, we evaluate ten LLMs using four safety levels: refusal, safe assistance, risky assistance with safety guidance, and fully risky assistance. Our results reveal greater vulnerability to education-specific risks and dynamic multi-turn interactions, while existing safety guardrails fail to adequately address these risks. EduZone advances LLM safety in education by providing an automated, scalable evaluation framework that supports the development and deployment of safer LLMs in K-12 education.

cs.AI↗

arXiv2026

Local statistics and average rank of genus $g$ hyperelliptic curves with a Weierstrass point

In this paper, we determine the probability that a genus $g$ hyperelliptic curve with a Weierstrass point over a number field has good reduction at a given prime of residue characteristic $>2g+1$. We also obtain analogous probability formulas for several other reduction types, including cases with positive toric or unipotent rank. As an application, assuming the Hasse--Weil conjecture and the generalized Riemann hypothesis, we derive an explicit upper bound for the average analytic rank of genus $g$ hyperelliptic curves with a Weierstrass point.

math.NT↗

arXiv2026

Pluralis v0.1: Towards a Multicultural, Multimodal, Multilingual Benchmark for AI Risk and Reliability

Current AI safety evaluation and benchmarking frameworks predominantly rely on Western-centric culture-agnostic defaults that mask critical regional laws, socio-linguistic nuances, and cultural taboos, leaving Vision-Language Models (VLMs) vulnerable in global deployments. We introduce Pluralis v0.1: a novel multimodal, multi-regional, and multilingual dataset built from a culture-first perspective. Spanning 6,448 prompts across six Asia-Pacific countries (Bangladesh, India, Korea, Pakistan, Singapore, Taiwan) and eight languages, Pluralis diverges from prior work by natively sourcing localized safety hazards rather than adapting Western datasets. Crucially, it introduces a multimodal evaluation paradigm: user text (e.g., "Should I gift this?") and an image referring to "this" (e.g., a clock) - both innocuous in isolation, but synergistically triggering specific legal or cultural violations. Pluralis disentangles universal safety violations from localized cultural appropriateness, establishing the latter as a first-class evaluation axis. To operationalize this, we present Judge-Pluralis, an agreement-gated LLM-as-a-Judge ensemble trained on examples classified in an empirically derived cultural taxonomy. Observing VLM behavior on a subset of the Pluralis surfaces recurring, locale-specific failure modes such as image misidentifications with downstream harm, missed item-context-locale interactions, and inadequate refusals. These failure modes vary systematically across locales and languages, exposing blind spots that globally averaged metrics conceal. Ultimately, Pluralis is not presented as a solved evaluation framework for cultural alignment, but rather as a first step and catalyst for future innovation. We call upon the research community to utilize this foundation to advance the science of multilingual, multicultural evaluation to better support AI cultural alignment globally.

cs.CL↗

arXiv2026

JuICE: A Benchmark for Evaluating LLM-Judge in Identifying Cultural Errors

As large language models (LLMs) are increasingly deployed to users around the world, they are integrated into everyday tasks across diverse cultural contexts, from drafting personal communications to brainstorming creative ideas. These tasks are inherently cultural: they require contextual appropriateness, symbolic resonance, and tacit cultural expectations that native speakers draw on instinctively, meaning that a response can be factually plausible yet unmistakably wrong to a local reader. Existing cultural benchmarks have treated culture as a flat set of facts via fact verification or norm entailment methods, and have adopted LLM-as-a-Judge without examining whether they can capture such thick cultural errors. To address this gap, we present JuICE (Benchmark for LLM-Judge in Identifying Cultural Errors), a multilingual dataset of 7,470 span-level annotations of cultural and linguistic errors in long-form LLM responses. It covers 1,050 query-response pairs from four countries (the United States, South Korea, Indonesia, and Bangladesh), in both English and their countries' main languages. Using JuICE, we find that even the strongest LLM-judge achieves only an F1 of 0.52 in the erroneous span detection task. Furthermore, LLM-judges consistently miss thick cultural errors that local residents readily identify. Our findings suggest that robust cultural evaluation must move beyond surface-level detection toward frameworks that account for the depth and situatedness of cultural meaning.

cs.CL↗

arXiv2026

Are they lovers or friends? Evaluating LLMs' Social Reasoning in English and Korean Dialogues

As LLMs are increasingly deployed in real-world interactions, their social reasoning in interpersonal communication becomes critical. To explore their capabilities, we introduce SCRIPTS, a 1.1k-dialogue dataset in English and Korean, sourced from movie scripts and propose a social reasoning task based on SCRIPTS that evaluates the capacity of LLMs to infer the social relationships (e.g., friends, lovers) between speakers in each dialogue. Evaluating nine models on our task, current LLMs achieve around 75--80% on the English dataset and 58--69% in Korean, and models predict an Unlikely relationship in 10--25% of responses in both languages. Furthermore, we find that thinking models and chain-of-thought prompting provide minimal benefits for social reasoning and occasionally amplify social biases. In sum, there are significant limitations in current LLMs' social reasoning capabilities, especially for Korean, highlighting the need for efforts to develop socially-aware LLMs across languages.

cs.CL↗

arXiv2026

Investigating Counterfactual Unfairness in LLMs towards Identities through Humor

Humor holds up a mirror to social perception: what we find funny often reflects who we are and how we judge others. When language models engage with humor, their reactions expose the social assumptions they have internalized from training data. In this paper, we investigate counterfactual unfairness through humor by observing how the model's responses change when we swap who speaks and who is addressed while holding other factors constant. Our framework spans three tasks: humor generation refusal, speaker intention inference, and relational/societal impact prediction, covering both identity-agnostic humor and identity-specific disparagement humor. We introduce interpretable bias metrics that capture asymmetric patterns under identity swaps. Experiments across state-of-the-art models reveal consistent relational disparities: jokes told by privileged speakers are refused up to 67.5% more often, judged as malicious 64.7% more frequently, and rated up to 1.5 points higher in social harm on a 5-point scale. These patterns highlight how sensitivity and stereotyping coexist in generative models, complicating efforts toward fairness and cultural alignment.

cs.CL↗

arXiv2026

Analyzing Process Data from Computer-Based Assessments: A Tutorial on Preprocessing, Feature Extraction, and Model-Based Inference

Computer-based assessments routinely generate detailed interaction logs -- commonly referred to as process data -- that record every action a respondent performs during task completion, yet systematic preprocessing guidance, integrated analytical workflows, and cross-method consistency checks remain scarce in the literature. This paper provides a unified, end-to-end analytical framework for analyzing process data from large-scale assessments -- covering the full pipeline from raw log preprocessing to model-based inference -- using the Programme for the International Assessment of Adult Competencies (PIAAC) Problem Solving in Technology-Rich Environments (PS-TRE) domain as an illustrative example. We first present a systematic preprocessing pipeline -- including timestamp correction, duplicate removal, action block consolidation, and LLM-assisted standardization -- that transforms raw event-level logs into analysis-ready action sequences. We then review and demonstrate two complementary families of analytical methods. The first consists of feature-based methods and their downstream applications, including descriptive process indicators, n-gram analysis with TF--IDF weighting, multidimensional scaling, and process data-informed differential item functioning (DIF) analysis. The second consists of model-based approaches, namely hidden Markov models and the subtask identification procedure. Empirical illustrations using the United States sample illustrate that n-gram-based behavioral clusters carry differential diagnostic information primarily among incorrect respondents, that multidimentionsl scaling-derived features comprehensively reconstruct observed behavioral variables, and that process-informed DIF analyses can identify and mitigate construct-irrelevant sources of group differences. Reproducible R code implementations are provided for all major techniques.

stat.AP↗

arXiv2026

Black Hole Properties of Type-1 Active Galactic Nuclei in the North Ecliptic Pole Wide Field: I. Mid-infrared Sources with Optical Counterparts

We present measurements of black hole (BH) properties of 861 Type-1 active galactic nuclei (AGNs) in the North Ecliptic Pole (NEP)-Wide field. These AGNs are detected in both optical and mid-infrared (MIR) surveys and are identified as Type-1 AGNs in optical spectroscopic surveys. By performing spectral energy distribution (SED) and line fitting, we obtained their MIR continuum luminosities ($L_{\rm MIR}$) as well as full width at half maximum (FWHM) values for the \ion{C}{4}, \ion{Mg}{2}, H$β$, and H$α$ lines. Using these measurements, we derived bolometric luminosities ($10^{43.20}$--$10^{47.27}~{\rm erg~s^{-1}}$), BH masses ($10^{7.29}$--$10^{9.67}$\,$M_{\odot}$), and Eddington ratios ($10^{-2.74}$--$10^{-0.08}$) for $\sim$450 objects over a wide redshift range ($z=0.09$--$4.71$). The use of $L_{\rm MIR}$ and FWHM values effectively alleviates the effects of dust extinction, enabling reliable estimates of BH properties even for dust-obscured AGNs. Moreover, we find that 34\,\% of the Type-1 AGNs in the NEP-Wide field are dust-obscured, and that their bolometric luminosities can be significantly underestimated without proper dust extinction correction. Our relatively extinction-free BH property estimates can (i) be combined with multi-wavelength data in the NEP-Wide field to facilitate diverse studies of AGN environments, number densities, host galaxies, and related topics, and (ii) serve as fiducial estimates for SPHEREx and other upcoming infrared (IR) spectroscopic missions covering the NEP-Wide field.

astro-ph.GA↗

arXiv2026

Arithmetic BF theory and the Cassels-Tate pairing

We give a systematic treatment of the arithmetic BF theory, introduced by Carlson and Kim. We observe that the Cassels-Tate pairing can be naturally interpreted as an arithmetic BF functional.

math.NT↗

arXiv2026

Goldfeld conjecture for non-hyperelliptic direction

Since the curve $y^2 = x^6+1$ has a large automorphism group, there exist twist families arising from non-hyperelliptic directions. In this paper, we give an explicit upper bound on the average analytic rank of such a family, assuming the generalized Riemann hypothesis for the $L$-functions. Also, we propose an analogue of the Goldfeld conjecture for the family following Katz--Sarnak philosophy.

math.NT↗

arXiv2025

Understanding EFL Learners' Code-Switching and Teachers' Pedagogical Approaches in LLM-Supported Speaking Practice

For English as a Foreign Language (EFL) learners, code-switching (CSW), or alternating between their native language and the target language (English), can lower anxiety and ease communication barriers. Large language models (LLMs), with their multilingual abilities, offer new opportunities to support CSW in speaking practice. Yet, the pedagogical design of LLM-based tutors remains underexplored. To this end, we conducted a six-week study of LLM-mediated speaking practice with 20 Korean EFL learners, alongside a qualitative study with nine English teachers who designed and refined responses to learner CSW. Findings show that learners used CSW not only to bridge lexical gaps but also to express cultural and emotional nuance, prompting teachers to employ selective interventions and dynamic scaffolding strategies. We conclude with design implications for bilingual LLM-powered tutors that leverage teachers' expertise to transform CSW into meaningful learning opportunities.

cs.HC↗

arXiv2025

World in a Frame: Understanding Culture Mixing as a New Challenge for Vision-Language Models

In a globalized world, cultural elements from diverse origins frequently appear together within a single visual scene. We refer to these as culture mixing scenarios, yet how Large Vision-Language Models (LVLMs) perceive them remains underexplored. We investigate culture mixing as a critical challenge for LVLMs and examine how current models behave when cultural items from multiple regions appear together. To systematically analyze these behaviors, we construct CultureMix, a food Visual Question Answering (VQA) benchmark with 23k diffusion-generated, human-verified culture mixing images across four subtasks: (1) food-only, (2) food+food, (3) food+background, and (4) food+food+background. Evaluating 10 LVLMs, we find consistent failures to preserve individual cultural identities in mixed settings. Models show strong background reliance, with accuracy dropping 14% when cultural backgrounds are added to food-only baselines, and they produce inconsistent predictions for identical foods across different contexts. To address these limitations, we explore three robustness strategies. We find supervised fine-tuning using a diverse culture mixing dataset substantially improve model consistency and reduce background sensitivity. We call for increased attention to culture mixing scenarios as a critical step toward developing LVLMs capable of operating reliably in culturally diverse real-world environments.

cs.CV↗

arXiv2025

One-Topic-Doesn't-Fit-All: Transcreating Reading Comprehension Test for Personalized Learning

Personalized learning has gained attention in English as a Foreign Language (EFL) education, where engagement and motivation play crucial roles in reading comprehension. We propose a novel approach to generating personalized English reading comprehension tests tailored to students' interests. We develop a structured content transcreation pipeline using OpenAI's gpt-4o, where we start with the RACE-C dataset, and generate new passages and multiple-choice reading comprehension questions that are linguistically similar to the original passages but semantically aligned with individual learners' interests. Our methodology integrates topic extraction, question classification based on Bloom's taxonomy, linguistic feature analysis, and content transcreation to enhance student engagement. We conduct a controlled experiment with EFL learners in South Korea to examine the impact of interest-aligned reading materials on comprehension and motivation. Our results show students learning with personalized reading passages demonstrate improved comprehension and motivation retention compared to those learning with non-personalized materials.

cs.CL↗

arXiv2025

The average analytic rank of elliptic curves with prescribed level structure

Assuming the Hasse--Weil conjecture and the generalized Riemann hypothesis for the $L$-functions of the elliptic curve, we give an upper bound of the average analytic rank of elliptic curves over the number field with a level structure such that the corresponding compactified moduli stack is representable by the projective line.

math.NT↗

arXiv2025

CrafterDojo: A Suite of Foundation Models for Building Open-Ended Embodied Agents in Crafter

Developing general-purpose embodied agents is a core challenge in AI. Minecraft provides rich complexity and internet-scale data, but its slow speed and engineering overhead make it unsuitable for rapid prototyping. Crafter offers a lightweight alternative that retains key challenges from Minecraft, yet its use has remained limited to narrow tasks due to the absence of foundation models that have driven progress in the Minecraft setting. In this paper, we present CrafterDojo, a suite of foundation models and tools that unlock the Crafter environment as a lightweight, prototyping-friendly, and Minecraft-like testbed for general-purpose embodied agent research. CrafterDojo addresses this by introducing CrafterVPT, CrafterCLIP, and CrafterSteve-1 for behavior priors, vision-language grounding, and instruction following, respectively. In addition, we provide toolkits for generating behavior and caption datasets (CrafterPlay and CrafterCaption), reference agent implementations, benchmark evaluations, and a complete open-source codebase.

cs.AI↗

arXiv2025

Diffusion Models Through a Global Lens: Are They Culturally Inclusive?

Text-to-image diffusion models have recently enabled the creation of visually compelling, detailed images from textual prompts. However, their ability to accurately represent various cultural nuances remains an open question. In our work, we introduce CultDiff benchmark, evaluating state-of-the-art diffusion models whether they can generate culturally specific images spanning ten countries. We show that these models often fail to generate cultural artifacts in architecture, clothing, and food, especially for underrepresented country regions, by conducting a fine-grained analysis of different similarity aspects, revealing significant disparities in cultural relevance, description fidelity, and realism compared to real-world reference images. With the collected human evaluations, we develop a neural-based image-image similarity metric, namely, CultDiff-S, to predict human judgment on real and generated images with cultural artifacts. Our work highlights the need for more inclusive generative AI systems and equitable dataset representation over a wide range of cultures.

cs.CV↗