Korean

Large language models (LLMs) are typically evaluated on static benchmarks, even though natural language constantly evolves through newly emerging words and meanings. Existing Korean benchmarks are centered on established vocabulary and therefore provide limited coverage of such recent lexical change, and their English-oriented design makes it difficult to assess the typological properties of Korean, in which content words combine productively with functional morphemes. In this paper, we introduce KoNeoBench, a benchmark for evaluating LLMs' understanding of Korean neologisms. KoNeoBench is built on 1,785 Korean neologisms attested in online news since 2020 and curated through expert lexicographic review. Each entry provides usage examples, word-formation analyses, and dictionary-style definitions. Based on this resource, we define four tasks and report results on recent models, together with a human baseline. Our experiments show that current LLMs exhibit clear limitations in recovering source components, distinguishing semantic categories, and generating accurate definitions. These results reveal specific aspects of recent Korean lexical change that remain challenging for current LLMs. KoNeoBench is available at https://github.com/bcmilab/ko-neobench/ .

High-quality 3D talking face datasets remain largely English- centric, and Korean 3D facial motion data are difficult to combine with standard English benchmarks because of differences in mesh topology, spatial scale, coordinate system, and temporal sampling. We present KoUniTalk, a lightweight articulation-centered Korean-English 3D talk- ing face benchmark that retargets VOCASET and the released Korean speech-based 3D talking face data to a shared mesh topology using de- formation transfer. Rather than proposing a new deformation-transfer algorithm or a full-head identity-preserving avatar dataset, KoUniTalk provides an identity-neutral canonical output space for controlled speech- driven facial articulation training and evaluation across English and Ko- rean. The unified template contains 1,176 vertices and focuses on the mouth and adjacent lower- and mid-face regions, reducing the output dimensionality from 15,069 and 72,147 dimensions to 3,528 dimensions, corresponding to 4.27-fold and 20.45-fold reductions compared with VO- CASET/FLAME and the original Korean mesh, respectively. To exam- ine whether retargeting preserves speech-relevant motion, we evaluate semantic mouth-landmark trajectories, including mouth opening, mouth width, aperture ratio, and mouth-opening dynamics. Since the official test set of the Korean dataset is not publicly released, we additionally define a subject-disjoint Korean benchmark split. The processed matched benchmark contains 22 speakers, 4,978 sequences, and 642,781 frames, enabling Korean-English cross-dataset evaluation of speech-driven 3D fa- cial animation models in a single compact articulation-template space. Source-reported inventory counts are listed separately from these pro- cessed counts

An AI assistant that calls tools makes two decisions on every request: which tool to invoke, and whether any available tool applies. In the usual design a single language model makes both, by emitting a call or by declining to emit one. On a device that has to answer without a server, the language model is what makes that design expensive, dominating both the latency and the memory of the router. The common alternative is to remove the model completely and rank the catalog of local actions with a retriever instead. That substitution is not symmetric across the two decisions. A retriever returns its highest-scoring candidate for every input and cannot signal that the catalog holds no valid action. Our earlier study found that constraining a decoder to a tool grammar repairs malformed output without improving the choice. What the substitution costs in each decision has not been measured. We evaluate the two decisions separately over 600 Korean and English requests and a catalog of 70 local actions. The router may also ask for a missing slot, reply, or delegate. Half the in-catalog requests reuse catalog vocabulary and half paraphrase it, separating lexical overlap from the action requested. Character 3-gram BM25 selects 162 of 164 lexically matched requests and 85 of 166 paraphrases. Restricting the candidate set to seven raises the paraphrase figure to a mean of 0.825 over five trials. No classifier over its score features separates in-catalog from out-of-catalog above 0.697 area under the curve, where the frozen encoder multilingual-e5-base reaches 0.806. Using that encoder for abstention alone keeps 376 of the requests local and misroutes 9 of the 150 needing delegation. Abstention, not selection, is where a neural component is required. A neural ranker improves every quality metric and is rejected on latency and memory rather than accuracy.

We post-train Qwen3.8-27B for Korean response style -- verbosity, list and markdown usage, discourse structure and register -- and measure two behaviours the objective never targets: abstention on ambiguous social questions in KoBBQ, where the benchmark-correct answer is UNKNOWN, and unprompted disclosure in securities guidance. Both move, and the changes are expressed primarily through the model's emission policy: how often it answers and how much it says. Matched target-form controls show that answer propensity depends on the training target, not the prompt set or recipe alone. Holding prompts, recipe, data volume and serving fixed and changing only the target text, three style seeds give positive answer-rate point estimates (mean +0.82 pp) and three neutral seeds negative ones (mean -1.53 pp); the observed seed ranges do not overlap and the means differ by 2.34 pp. A length-matched arm lies between them, and a fourth arm that stays short while preserving hedging is unstable across seeds, so which feature of the form is responsible is unresolved. For absolute stereotyped exposure the decomposition into an answer-propensity term and a conditional-composition term is an algebraic identity, not a finding; its empirical content is where the movement went. Across the trained checkpoints the changes are dominated by answer propensity while the composition term stays small, and because that term is evaluated on treatment-dependent answered subsets we do not read it as evidence about latent preference. Two measurement results follow. A between-arm contrast in conditional stereotyped share does not identify a change in conditional content preference when answer status is treatment-dependent. And agreement between two rule detectors for the same construct runs from 0.44 to 0.99 depending on which checkpoint produced the text -- observable without any reference labels.

Automatic Speech Recognition (ASR) technology is fundamental to customer service automation and large-scale transcription. However, even advanced ASR models exhibit inevitable errors in complex real-world environments such as call center conversations. When privacy restrictions preclude audio access, error correction must rely on text-based post-editing. Existing text-only approaches face significant challenges in low-resource languages, mainly due to a critical scarcity of annotated corpora and tailored correction methodologies. For Korean, this resource gap is particularly pronounced, as existing resources are predominantly designed for ASR training rather than text-based error correction. To address this, we introduce DasanCallDial, the first large-scale Korean benchmark dataset specifically curated for dialogue-level ASR error correction. Derived from genuine call center interactions, it comprises 1,974 dialogues with 115,460 utterances. Leveraging this resource, we propose Detector-Gated Contextual Span Correction (DCSC), a text-only post-editing framework for error-sparse Korean speech recognition transcripts. DCSC combines an encoder-based detector that first performs token-level error detection, followed by a language model-based corrector trained to rectify fine-grained span-level errors. Additionally, we employ dialogue-level context augmentation to enable the model to leverage discourse history for disambiguation. By employing multi-level granularity, our method achieves state-of-the-art performance, effectively overcoming the limitations of general LLMs in low-resource settings.

Multi-agent LLM deliberation has been explored as a scalable way to simulate public deliberation. For such simulations to be informative, persona agents should reflect population opinion patterns and interaction should shape their conclusions. We evaluate whether LLM-based deliberation can meet these two conditions using census-grounded Korean personas debating real policy questions benchmarked against national surveys. Persona agents do not reliably reproduce population opinion patterns: responses are often far more concentrated and frequently reverse demographic differences in the human data. Deliberations nonetheless produce reasoned, reciprocal, and varied arguments alongside substantial stance movement. Yet much of this movement does not require peer exchange: sealed-monologue agents change position at similar rates and reach nearly the same final balance as full debates, while groups initialized with very different positions often converge to similar endpoints. Anchoring population-informed starting positions, meanwhile, sharply suppresses updating. Thus, population representation, argument generation, and interaction-driven opinion change do not necessarily go together. The simulations readily surface arguments on both sides, though whether they capture the diversity of human perspectives remains untested, leaving open a promising role for argument surfacing even as population simulation requires further validation.

Excess vocabulary, a word's frequency above its pre-2023 trend, is how the change in scholarly English after 2022 has been measured. We adapt it to Korean with morphological units on 398,296 KCI abstracts (2018-August 2026), with 47,165 Vietnamese abstracts for comparison. Placebo floors are 0.1-2.2 points for the single-word statistic and at most 2.9 for the re-selected split-half set statistic. Korean abstracts show nothing in 2023, onset in late 2024, a rise through 2025 flattening in mid-2026: sisahada "suggest" appears in 21.4% of 2026 abstracts against 5.3% expected; plain verbs like araboda "look into" fall to a quarter of trend. Under stated assumptions the single-word conditional lower bound on LLM-processed abstracts is 3.5%, 10.5% and 16.1% for 2024-2026 and a split-half set bound 7.8%, 20.6% and 33.0%. Holzwarth et al.'s estimator under the same discipline gives 41.9% and 72.1% for 2025-2026. Subject-matter controls reduce but do not remove it: restricting the set to lemmas three language-model annotators all call style leaves 14.7 of the 33.0 points, and pairing each 2026 abstract with its journal's closest base-period abstract leaves 34.1. Tested translation routes do not explain it: the surface marks of translated Korean fall as the markers rise. In the same articles' English abstracts the excess appears a year earlier; where the English side carries none, the Korean shift persists at 30 to 66% of the rate where it does. Control abstracts from three providers reproduce the rising words, with marker turnover consistent with model generations; implied prevalences are scenario-dependent.

Korean introduces an additional typographical perturbation level not captured by ordinary character-level edit models: because syllable blocks are internally composed of sub-character units called jamo, keyboard-level errors can occur within a syllable, either producing a valid but semantically altered character or exposing raw jamo on the surface. Both outcomes disrupt sub-word tokenization and are not reliably corrected by existing grammatical error correction pipelines, leaving LLMs directly exposed to corrupted inputs. To quantify this vulnerability, we apply five jamo-level perturbation types to the KMMLU benchmark and evaluate four language models, finding that accuracy declines monotonically with perturbation intensity and that parameter scaling does not confer robustness against intra-syllabic noise. We further show that typo-corrupted inputs induce a distinct shift in internal representations that is not reducible to ordinary answer incorrectness, and that a simple linear probe trained on these representations detects unseen perturbation types with high AUROC. Motivated by this signal, we propose Typo-Aware Chain-of-Thought (TACoT), which routes inputs to chain-of-thought inference only when the probe detects a likely typo, recovering a substantial portion of the CoT accuracy gain at a fraction of the inference cost.

We introduce A.X K2, a 688B-parameter Mixture-of-Experts (MoE) language model trained from scratch as a high-performance foundation for \emph{agentic} applications. Trained on approximately 8.5T tokens---fewer than its predecessor, A.X K1---on a smaller but higher-quality mixture with substantially expanded agentic and software-engineering data, it nonetheless improves over A.X K1 across the board, by over 30 percentage points on some benchmarks, reflecting large gains in token efficiency. To support long contexts efficiently, we introduce Sparse Gated Attention (SGA), which combines sparse attention with gated attention, and adopt Gated Norm (GN) to stabilize large-scale training. SGA is trained natively at 128K through a \emph{sparse} indexer warmup that optimizes the indexer against its own sparse top-$k$ selection rather than the dense attention distribution, making adaptation markedly cheaper: each query reads only 2,048 positions, yet long-context quality is unchanged and A.X K2 scores 94.6 on RULER out to 256K. The outlier suppression of GN in turn keeps 4-bit NVFP4 serving within one point of FP8 accuracy. A simple yet effective Think-Fusion recipe further lets users switch between thinking and non-thinking modes within a single unified model. Extensive evaluations show that A.X K2 performs competitively against strong open-weight baselines, matching or exceeding them on math and Korean-language benchmarks.

Popular music circulates globally while being locally reinterpreted, yet this process of cross-cultural style diffusion has rarely been quantified. We propose an era-classification framework for measuring temporal alignment between chart cultures. CNN classifiers trained from scratch on Billboard Hot 100 audio are applied to Korean Melon chart songs. Korean chart songs from the 1960s through the 1980s are consistently inferred as belonging to earlier Billboard eras, by a median of about four to five years, while the same models remain unbiased on held-out Billboard audio. The offset then halves at the 1990s, to roughly two to three years, and holds there through the 2000s. Reverse inference shows a complementary narrowing, and the pattern holds across architectures and seeds. We interpret these results as reflecting how globally circulating pop styles were locally adopted and progressively synchronized. The framework can be applied to other pairs of chart cultures beyond the US-Korea case examined here.

Computational analysis of music often relies on discrete representations, yet many musical traditions are organized around continuous pitch movement that resists segmentation into note-like units. For such traditions, the discrete units that analysis would build on are not given in advance. We address this gap by learning a vocabulary of local pitch-contour patterns directly from unlabeled audio, using a VQ-VAE that quantizes fixed-length contour segments into a finite codebook. To make the learned tokens stable across segmentation positions and small variations in timing and pitch range, we train the model with a reconstruction objective evaluated under the best alignment among a set of candidate temporal and pitch-domain transformations. Applied to Korean traditional music, the learned tokens recover information about expert-defined sigimsae categories without supervision, and in pansori individual tokens align with the two principal modes, Gyemyeonjo and Ujo, supporting their use as units for corpus-level analysis of contour-centric traditions.

Pansori is a traditional Korean vocal genre whose mode system (jo) is defined not by scale alone but by the entanglement of pitch collection, microtonal ornament (sigimsae), and vocal timbre. In this study, we introduce a 46-hour frame-level pansori mode annotation, expert-labeled across all five canonical batang, and evaluate four complementary input representations (mel spectrogram, F0 contour, MIDI piano roll, and a multi-cultural SSL encoder) under two split strategies designed to detect shortcut learning. Across the three well-represented modes, performance degrades by only 2.1--3.6 points of F1 when entire works are held out, indicating that the models learn mode-relevant features rather than memorizing repertoire. Per-class results further show that source separation removes the percussion cue on which changjo depends, and that generic multi-cultural pre-training fails specifically on the Ujo--Gyemyeonjo distinction. Qualitative analysis of cross-modal disagreement recovers musicologically documented phenomena and agrees with published score-based analyses of modern changjak pansori.

We introduce NOLLI, a procedurally generated English-Korean puzzle benchmark designed to diagnose where Korean performance gaps arise. It comprises 15 puzzle types (25 tasks; 7,500 items), with every instance seed-regenerable, verified to have a unique solution, and scored deterministically. Rather than equating harder with bigger, we calibrate difficulty behaviorally, tuning each generator until a fixed reference model lands in target accuracy bands. Its three-level design combines matched direct translations, script adaptations over Hangul jamo (sub-syllabic letters), and Korean-only tasks grounded in Korean culture or orthography. We evaluate 15 frontier, open-weight, and Korean-developed models; among the 12 above a 3% overall-accuracy floor, matched English-Korean accuracy is statistically equivalent within a +/- 10 pp margin (TOST), suggesting little cost from presentation language alone. Writing-system-intensive tasks show sharper gaps: Korean Cipher falls behind English by up to 68.7 pp, whereas Cryptarithmetic over the same jamo shows no systematic penalty, and Jamo Composition accuracy predicts Korean Cipher accuracy. These contrasts are diagnostic rather than causal, consistent with difficulty in multi-step sub-syllabic execution. Korean-only tasks separate rule-application deficits, which vary in sign, from a Kinship deficit positive in all 12. Finally, a salient size measure fails to grow from Easy to Hard in 7 of 15 types, making structural size an unreliable proxy for empirical difficulty.

Document packets, multiple documents concatenated into a single file, are common in government and administrative workflows, yet splitting them into their constituent documents is difficult, especially for low-resource languages. We introduce Khondo (Bangla for split/segment), the first benchmark for document packet splitting on Bangladeshi government forms. Unlike prior English and OCR-text-based datasets, Khondo is bilingual (Bangla--English) and vision-native; where models operate directly on page images. It spans five concatenation schemes, from sequential to fully shuffled, across 14 administrative domains, with ground-truth boundaries, domain types, and page order. Zero-shot evaluation of MLLMs shows they cluster pages into their source documents fairly well but struggle in restoring the original page order once shuffled. To isolate what drives this difficulty, we run two controlled analyses, varying the prompt instruction and then the packet language. Both primarily affect ordering rather than clustering: (a) explicit page-order instructions are necessary but insufficient, and (b) English packets are ordered more reliably than Bangla, making page arrangement the dominant challenge and language a secondary but consistent factor. Khondo establishes page-order reconstruction as a key open problem in vision-based, low-resource document understanding, and provides a controlled benchmark for measuring progress toward solving it. Our dataset and code is available at https://huggingface.co/datasets/Mausul/khondo

We present Solar Open 2, a 250B-A15B Mixture-of-Experts language model built for long-horizon agentic tasks, scaled up from Solar Open 1 (Solar Open 100B). To hold entire agent trajectories in a single context, Solar Open 2 reaches a 1M-token window through a hybrid attention stack that interleaves one softmax layer among every three linear-attention layers, using no positional encoding and a gated delta rule extended to negative eigenvalues. To train at this scale under a fixed compute budget, we make training efficient in two ways: a stronger starting point, and higher-value data. For the starting point, we initialize Solar Open 2 from Solar Open 1, transferring the 5.69B-parameter shared skeleton that survives the architectural change and learning everything else through full pre-training. For the data, we curate for value per token: quality- and rarity-aware data curation and mixture-ratio optimization refine a 20T pool into a 10T mixture that, at equal token budget, outperforms the Solar Open 1 recipe. To build its agent skills, we train twelve domain specialists across purpose-built scenarios, then consolidate them into a single model by Multi-teacher On-Policy Distillation (MOPD). Against comparably sized open-weight models on English benchmarks, Solar Open 2 leads on MMLU-Pro, LiveCodeBench, and the APEX-Agents agentic suite, and stays competitive with the strongest (DeepSeek-V4-Flash and MiMo-V2.5) elsewhere. On Korean benchmarks, Solar Open 2 records the highest average of any model compared, including fast-tier closed APIs, and on Ko-GDPval, an in-house Korean officework-agent benchmark, it is competitive with DeepSeek-V4-Pro (1.6T) at less than a sixth of its size.

The rise of generative AI (GenAI) in higher education has prompted urgent debates surrounding academic integrity and ethical use. This study examines cross-cultural differences in student perceptions of GenAI use, comparing responses from students at Canadian and South Korean universities. Using a scenario-based survey administered in Fall 2024, we analyzed how students judged the ethicality and rule compliance of AI-assisted coding practices. Results reveal that Canadian students were consistently more likely to perceive the use of GenAI as both unethical and against institutional policies compared to Korean students, despite functionally identical institutional policies. Statistical analysis, including Mann-Whitney U tests and correlation coefficients, demonstrated significant differences across nearly all scenarios. Analysis of the factors used in generating scenarios indicated that the amount of AI-generated code incorporated into assignments most strongly influenced ethical judgments. Findings were interpreted through Hofstede's cultural dimensions framework, suggesting that cultural factors such as power distance, individualism, and uncertainty avoidance significantly shape students' ethical reasoning regarding GenAI. Our results contribute to the growing body of evidence emphasizing that equitable AI integration in education must be culturally responsive, taking into account diverse conceptions of academic integrity. We advocate for the development of nuanced AI-use guidelines that are sensitive to local cultural contexts while upholding fundamental principles of academic honesty. This study highlights the need for ongoing cross-cultural research to inform ethical AI policies and support responsible GenAI use in global higher education settings.

AI_LectureNote is a historical, readability-oriented post-ASR workflow for Korean-English medical lectures. It rewrites speech-to-text output into study transcripts while restoring Latin-script medical terms rather than Korean phonetic transliterations. We retrospectively evaluate the workflow on four author-recorded lectures across five conditions. In this pilot, post-processing raised the macro English-script rendering rate from 0.39 to 0.71 on the whisper-1 path and from 0.26 to 0.65 when applied to 3-minute chunked gpt-4o-transcribe output. However, English-script rendering did not imply semantic faithfulness: the two post-processed conditions showed semantic drift in 34 and 36 of 282 reference sentences and polarity failures in 11 and 13 of 101 polarity-cue rows. A descriptive cross-input comparison suggested different candidate failure patterns: polarity-failure sets overlapped more strongly across front-ends (Jaccard 0.60; 9 shared of 15 unioned failures) than general semantic-drift sets (Jaccard 0.23; 13 shared of 57 unioned drifts). This single-annotator pilot documents concrete failure modes rather than population rates and supports evaluating surface accuracy, term-script rendering, chunk-level script consistency, and medical-meaning preservation separately.

The Korean adnominal ending \texttt{ETM} occurs in diverse noun-modifying constructions, including relative-clause-like modifiers, adjectival and copular forms, bound-noun constructions, and lexicalized expressions. This paper argues that \texttt{ETM} is not a direct marker of relative-clause structure, but a morphological exponent shared by several adnominal constructions. We propose a corpus-based typology that distinguishes these constructions using predicate type, auxiliary structure, argument-structural compatibility, head-noun restriction, and lexicalized patterns. We operationalize the typology as a construction-sensitive annotation layer for the KLUE dependency treebank, implemented through an ordered rule-based procedure and evaluated by manual validation. Productive relative-clause-like uses account for 39.4% of the analyzed instances; the remainder consists mainly of adjectival, copular, bound-nominal, modal, temporal, and collocational constructions. The findings show that Korean relative-clause-like modification cannot be identified from adnominal morphology alone.

We present LuckyStar 111B, a 111B-parameter hybrid reasoning model developed through a collaboration between Cohere and LG CNS for Korean-English enterprise agents under practical memory and serving constraints. The model trains from Cohere's fully post-trained Command A model rather than a new pretraining run, and uses preamble conditioning to switch between concise non-reasoning behavior and longer tool-oriented reasoning. We study four choices for scaling tool-using agents efficiently: multilingual supervised fine-tuning, reinforcement learning with verifiable rewards for multi-step tool-use tasks, language-consistency rewards for Korean user-facing responses, and 4-bit quantization for single-GPU serving. The adapted model improves mathematical reasoning, function calling, and agentic natural-language-to-SQL (NL2SQL) performance while preserving general Korean and English instruction-following quality. These results provide a practical recipe and failure-mode analysis for adapting post-trained multilingual models to verifiable agentic workflows under memory-constrained deployment.

Urban building change detection from bi-temporal aerial imagery is important for redevelopment monitoring, infrastructure management, and unauthorized-construction screening, but Korean urban scenes remain difficult because changed regions are often sparse, appearance varies strongly between acquisition dates, and useful outputs must follow building footprints rather than coarse blobs. This paper presents UrbanCDNet, a task specific Siamese CNN that combines appearance-robust multi-cue comparison, alignment-aware middle-scale differencing, lightweight context refinement, scene calibration, and auxiliary boundary supervision. Experiments use a corrected AIHub-based Korean benchmark with 3,998 training, 503 validation, and 499 test pairs, and report changed-class precision, recall, F1, and IoU. On the locked test split, UrbanCDNet achieves 0.7335 precision, 0.7696 recall, 0.7511 F1, and 0.6014 IoU, outperforming a strong Siamese U-Net baseline (0.7108 F1, 0.5514 IoU) and the strongest external competitor, ChangeFormer-MIT-B0 (0.7107 F1, 0.5512 IoU). Additional diagnostic slicing shows that the gain is concentrated in the operating regimes that motivated the design: on the sparse-change subset with less than 5% changed area, F1 improves from 0.4765 to 0.6175, and on the high photometric-gap subset it improves from 0.6349 to 0.7285. Boundary F1 at 3-pixel tolerance rises from 0.3445 to 0.4447, while object F1 at IoU 0.3 rises from 0.0690 to 0.2258. These results indicate that, on this Korean benchmark, task-shaped temporal comparison and boundary-aware supervision matter more than generic model scale alone

Large pretrained text-to-speech (TTS) models sound almost human for well-resourced languages, but much worse for languages that are rare in their training data. We study this quality gap for Khmer and Korean using VoxCPM2, a 2.4B-parameter, tokenizer-free TTS model that joins a MiniCPM-4 language-model backbone with a flow-matching diffusion decoder. We build one shared, language-tagged corpus of about 26 hours and adapt VoxCPM2 with a single Low-Rank Adaptation (LoRA) adapter, trained on both languages at once and added to both the language model and the decoder. The adapter is zero-initialized, so training starts exactly at the original (zero-shot) model. In native-speaker listening tests, the Khmer Mean Opinion Score (MOS) rises from 3.85 to 4.23 with the best adapter (rank 64), a highly significant gain (paired Wilcoxon test, p<0.001), while training only 0.19 to 3.03 percent of the parameters. The automatic loss and the human ratings, however, disagree on the best rank: validation loss is lowest at rank 128, yet MOS peaks at rank 64. The same adapter brings no gain for Korean, a language the base model already handles well, and at a high rank it even degrades quality. Adaptation therefore helps mainly where the base model is genuinely weak.

Coordinated foreign influence operations pose a growing threat to online platforms, but detecting state-linked troll activity and tracking its evolution remain challenging. This paper presents an explainable machine learning framework for theory-guided detection and longitudinal analysis of suspected trolling within Korean online news comment sections. Our hierarchical model classifies comments along three dimensions central to influence campaigns: foreign origin, moral-emotional framing, and target country. To support explainability, it also extracts brief span-level textual evidence that provides human-interpretable rationales. We apply the approach to 112M South Korean news comments authored by 4M users over nearly 20 years, identifying 23,998 accounts exhibiting behavior consistent with coordinated manipulation. Analyzing these accounts, we find that they predominantly rely on morally condemning rhetoric rather than direct promotion of foreign-aligned narratives; this rhetoric receives significantly higher user engagement. Among the highest-engagement comments, the moral condemnation most frequently targets domestic political figures (e.g., presidents or party leaders) on both the left and the right, potentially amplifying polarization. Our framework supports transparent platform governance through explainable, evidence-based moderation. These observed rhetorical and engagement patterns can inform how platforms and observatories prioritize defenses and intervene before harmful narrative-target combinations achieve widespread reach.

Political bias in large language models (LLMs) is increasingly significant, but difficult to measure reproducibly across political and linguistic contexts. We introduce Polar, a 4,026-instance multiple-choice benchmark that measures political bias through option-level likelihoods rather than prompt-based generation. Polar covers two ideological axes and eight issue categories derived from the Manifesto Project, and evaluates models in parallel across U.S. and South Korean political contexts. Across 38 LLMs, measured bias varies systematically with political context, issue category, model group, and presentation language. All models lean left-progressive on U.S. political content, but show more centered and mixed patterns on South Korean content. Translation experiments further show that presentation language alone can shift measured bias. These findings highlight the need for multilingual and cross-contextual evaluation of political bias in LLMs.

Math reasoning benchmarks have proliferated, yet most lack a per-item difficulty signal grounded in actual human performance. We introduce KCSAT-ML, a decade (2014-2025) of Korean College Scholastic Ability Test (KCSAT; Suneung) mathematics: 664 problems with a 339-item core set carrying official per-item error rates from nationwide cohorts of hundreds of thousands of examinees. We pair the benchmark with Difficulty-aligned Reasoning Gain (DRG): a score-orthogonal metric that asks whether a model's mistakes concentrate on the items humans found hard, or on items humans found easy. Together they expose, across a wide range of VLMs (and LLMs via OCR), three patterns: (i) low-budget accuracy collapses on the high-human-error tail at every model size; (ii) test-time scaling (TTS) raises token use roughly linearly with cohort error rate, while accuracy gains follow a non-monotonic curve; (iii) within a single family, TTS flips between anti-scaling on the hardest items and overthinking on easier ones -- two faces of the same alignment failure. On DRG, models with near-identical accuracy can sit at near-opposite values: one model gets wrong what humans also find hard, while another solves the hardest items yet fails on items humans find easy -- a contrast that aggregate accuracy hides. Our code and dataset builder will be open-sourced at https://github.com/naver-ai/KCSAT-ML.

Speech sound disorders affect approximately 44% of Korean pediatric communication disorder cases, yet automated assessment tools for Korean toddler speech remain underdeveloped. This paper presents an end-to-end pipeline for automated pronunciation evaluation of Korean toddler speech, combining neural speaker diarization with self-supervised speech representation learning. We introduce a novel IRB-approved corpus of 53 recordings from Korean-speaking children aged 2-5 years. A subset of 53 subjects was annotated by three independent reviewers, yielding 1,190 consonant and 748 vowel word-level binary correctness labels. We evaluate three diarization models, finding that NeMo SortFormer achieves 88.69% speaker count accuracy and 33.04% diarization error rate (DER) owing to its arrival-time-sorted transformer architecture, which handles the acoustic confound between young female caregivers exhibiting aegyo and toddler speech. For pronunciation scoring, we compare three self-supervised learning (SSL) backbones across multiple pooling strategies. A cross-model ensemble routing consonant prediction to HuBERT-large and vowel prediction to WavLM-large achieves balanced accuracies of 0.720 and 0.845, with a mean of 0.782.

Cultural-aspect work on large language models is dominated by a negative target: which outputs to suppress. We argue that a constructive counterpart is also needed, a working definition of what a culturally coherent response is rather than only what it must avoid, and instantiate it for Korean. We design an alignment-data pipeline around a prompt-based LLM seed generator that expands a Korean harm taxonomy, with a Korean-culturally-adapted safe-response policy at its centre: a per-category guideline grounded in Korean legal frameworks, social norms, and interpretive conventions, against which three frontier models each produce a candidate response. DPO fine-tuning on the resulting triplets improves the Korean cultural safe rate across six open-weight LLMs while causing no large degradation on Korean general-capability benchmarks, and qualitative outputs show fine-tuned models naming Korean statutes and institutional procedures and, where appropriate, supplying constructive Korean-context information alongside refusal.

T2I models cannot effectively capture sentiment from various types of text, including diaries, as they primarily focus on visual object-related patterns rather than contextual emotional understanding. This paper proposes an emotion-aware text-to-image pipeline that generates children's hand drawing style images from short Korean diary entries. The proposed pipeline employs Qwen3-8B for recognising implicit sentiment from short diaries, and Stable Diffusion 3.5 Medium fine-tuned with LoRA on children's drawing images with emotion-based trigger words for image generation. Additionally, this paper presents experiments examining the effect of emotion trigger words on generated images and discusses the limitations of CLIP Score as an evaluation metric for emotion-aware image generation.

Frontier model evaluations are shifting from foundational capabilities (e.g., instruction following and reasoning) toward compositional, agentic ones, but Korean agentic benchmarks remain scarce. We introduce K-BrowseComp, a web-browsing agent benchmark grounded in Korean contexts, consisting of 400 problems. The 300-problem K-BrowseComp-Verified subset is manually constructed and validated by native Korean speakers. On this subset, frontier LLMs, including GPT-5.5, DeepSeek-V4-Pro, and GLM-5.1, reach only 30.00--45.67%, a substantial drop from BrowseComp, while Korean LLMs released through Korea's Proprietary AI Foundation Model program obtain only 0.00--10.33%. We further construct a 100-problem synthetic split using hard few-shot exemplars and failure-mode-targeted generation to exploit the asymmetry between solving and creating web browsing problems. On the adversarially filtered synthetic diagnostic split, the strongest model reaches only 26.00%, and we report this split separately as a targeted stress test. We publicly release our data and code.

We audit six large language models (LLMs) for gender stereotyping across English, Korean, Chinese, and Japanese. Three were developed primarily for English-language use (Claude, GPT, Gemini) and three for East Asian use (DeepSeek, Syn-Pro, HyperCLOVA X). We adopt the HEXACO-100 personality inventory and anchor each model against a cross-cultural human dataset spanning 48 countries to ask not whether LLMs are biased, but how far their gender attributions drift from the populations they are deployed among. Our findings show that their stereotyping spans a range roughly 2.5 times wider than the entire cross-country range found in humans, and the effect can compound across languages. One English-centric model, prompted in Korean, reached 5 times the local baseline, even when the prompt stated the candidate had already been hired, which often dampens human stereotyping. To characterize such behaviors without ranking them, we introduce a four-pattern framework -- concordance, suppression, reorganization, and amplification -- across 24 (model x language) cells. Item-level analysis reveals that translation does not just rescale stereotypes, but changes the attributes tied to it, hiding significant rearrangement under the surface while appearing well-calibrated. Our results ultimately suggest that no single debiasing pipeline is likely to address bias evenly across linguistic boundaries.

E-learning systems should deliver contents that reflect various phenomena of the language as it is used. In addition to formal Korean, e-learning systems that would include real-world Korean expressions such as those in web documents, mobile text messages, or twitter posts, would be useful to high-level learners. We construct two types of corpora: one is made of formal documents like online news articles; the other is made of informal documents like customer reviews about new products in web blogs. By comparing these corpora, we show how expressions differ in these two types of corpora. We survey the main characteristics of the informal corpus. Given that a significant proportion of text is informal, we propose Local Grammar Graphs (LGG) as an appropriate model to treat them effectively in Korean e-learning systems.

Large Language Models (LLMs) have advanced financial automation through Retrieval-Augmented Generation (RAG), yet hallucinations remain a critical barrier to deployment in high-stakes environments. Existing benchmarks focus on single-turn, English-centric tasks, leaving the multi-turn dynamics and linguistic-regulatory nuances of the Korean financial domain unaddressed. We introduce K-FinHallu, the first benchmark for hallucination detection in multi-turn Korean financial RAG. We construct multi-turn dialogues from authentic Korean financial documents and inject hallucinations under a proposed hierarchical taxonomy based on context answerability that explicitly accounts for justified abstention. Benchmarking frontier and open-source LLMs as hallucination detectors, we find that even the strongest models struggle with fine-grained financial diagnostics and refusal behavior. While fine-tuning an 8B model on our training split yields performance competitive with frontier LLMs, justified abstention remains the weakest axis across all evaluated models.

Multimodal Large Language Models (MLLMs) exacerbate safety risks by introducing vulnerabilities across multiple modalities, such as language and vision. Current MLLM safety evaluation tools, however, suffer from major limitations: 1) English-centric dataset construction, and 2) a focus on generic risks that are not tied to local cultural contexts. This paper introduces KSAFE-MM, a benchmark for Korean multimodal safety evaluation that covers both general safety risks and culture-specific vulnerabilities. KSAFE-MM consists of two parts, KSAFE-MM-G and KSAFE-MM-C. KSAFE-MM-G evaluates globally shared risks in Korean contexts through linguistic contextualization, which transforms generic safety queries into contextually grounded multimodal samples. KSAFE-MM-C targets culture-dependent MLLM safety vulnerabilities using localized visual queries derived from real-world contexts. It pairs these visual queries with jailbreak-style textual queries to cover multimodal safety risks involving cultural visual cues and malicious textual intent. Together, these components provide a general-to-local construction pipeline for evaluating both globally shared safety risks and culture-specific vulnerabilities. We evaluate 12 state-of-the-art MLLMs on KSAFE-MM and reveal that models exhibit greater vulnerability to culturally grounded attacks than to generic ones. Notably, jailbreaking strategies substantially amplify attack success rates, with ProgramExecution yielding up to 74.2% ASR compared to 13.4% for standard queries. Furthermore, we identify a systematic trade-off between safety and over-refusal, where models achieving low ASR tend to exhibit excessive refusal behavior on benign queries. These findings highlight the urgent need for culturally grounded safety evaluation beyond English-centric benchmarks.

Speech language models (SpeechLMs) have achieved substantial progress by extending large language models (LLMs) to the speech modality. However, SpeechLM evaluation remains heavily centered on English, limiting reliable assessment of multilingual speech capabilities. Straightforward benchmark transfer through ASR, translation, normalization, and TTS can corrupt language-specific instructions, answer constraints, and spoken forms; for audio understanding, transferring source-language audio also fails to preserve target-language speaker attributes, accents, and paralinguistic properties. To address these limitations, we propose two human-agent benchmark-construction frameworks: one transfers source-language SpokenQA benchmarks into target-language SpokenQA benchmarks, and the other converts target-language ASR corpora into audio understanding benchmarks using transcriptions and speaker metadata. Using these frameworks, we construct and publicly release three Korean speech benchmarks: KVoiceBench and KOpenAudioBench for Korean SpokenQA, and KMMAU for Korean audio understanding, comprising 12,345 samples in total. We evaluate eight recent SpeechLMs and find that English-Korean performance gaps vary substantially across models and task families, and that SpokenQA and audio understanding rankings diverge, revealing complementary weaknesses invisible to English-only evaluation.

Korean folk painting (minhwa) is built from a small vocabulary of auspicious symbols, a tiger for protection, a pair of birds for marital harmony, a peony for wealth, that recur across many of its painted genres. This suggests an obvious computational approach, identify which symbols appear in a painting and read the genre from the inventory. Working with a public corpus that pairs whole paintings, eight-field bilingual curatorial captions, and a separate set of expert object crops, we find that this approach does not work. A model given only a list of which symbols a painting contains predicts the genre far worse than a model that fuses the image with the curatorial text, and forcing the genre representation to be object-grounded actively hurts accuracy. The visual evidence on which the genre prediction rests is nonetheless localized and inspectable. A leakage-safe object evidence map projected from a part-level detector is spatially faithful to where curators isolated symbolic objects and to a patch-based surrogate's own gradient saliency. We name this configuration a faithful-but-insufficient dissociation. The part-level explanation is honest about what the part-level model sees, yet the genre target turns on how symbols are arranged rather than on which ones appear. The same lens separates a content label that survives transfer to held-out source institutions, genre, from a style label that does not, era, a prediction we confirm on two further labels in the corpus. We release the multimodal system, a worked-example reading of one painting's evidence map against its catalogue, and a set of evaluation cautions that recur in long-tailed heritage collections.

Lexical gaps are words that do not exist in certain languages. They pose challenges for building multilingual lexical resources, for machine translation, and for cross-lingual transfer. Existing lexical gap detection relies on human judgments or fixed conceptual taxonomies. We propose a data-driven framework for identifying cross-lingual lexical gaps. We extracted contextualized embeddings from Korean-English bilingual LLMs for Korean-to-English and English-to-Korean translation pairs. Combinations of LLMs, embedding types, dimensionality, and orthogonal transformations across 100 train-test splits yielded 4000 distinct embedding spaces in each source language. In each space, we computed the semantic similarity between each source word and its nearest neighbor in the target language, and compared their distribution for gap words versus non-gap words. In 94% (Korean-to-English) and 97% (English-to-Korean) of embedding spaces, gap words showed weaker cross-lingual semantic alignment than non-gap words. Logistic classifiers trained on unaligned embedding spaces can reliably separate gap words from non-gap words, achieving AUCs of 0.81 (Korean-to-English) and 0.76 (English-to-Korean) and retrieving 18/19 Korean and 26/27 English gap words. This approach provides a language-agnostic and taxonomy-free method for scalable lexical gap identification.

While Large Language Models (LLMs) are increasingly integrated into in-vehicle conversational systems, identifying the optimal model remains challenging due to the lack of domain-specific evaluation standards tailored to real-world deployment requirements. In this paper, we propose a novel evaluation framework for in-vehicle assistants, with a particular focus on Korean-language localization. Our empirical analysis reveals notable patterns in model behavior. First, fine-grained Korean honorific control remains unstable in current LLMs, indicating that precise speech-level realization must be explicitly evaluated in localization settings. Second, models exhibit weaker performance in strategic conversational metrics like clarification and proactivity. Our analysis suggests this stems from the inherent subjective complexity of these tasks, where our framework adopts a conservative evaluation stance to prioritize reliability. Together, our findings underscore that automotive AI must move beyond general competence toward precise linguistic tailoring and reliable, safety-oriented interaction management.

We analyze how automatic speech recognition (ASR) errors propagate through ASR-LLM cascades in Korean spoken question answering (SQA), focusing on downstream semantic failures that conventional ASR metrics cannot fully capture. Our analysis shows that the relative downstream degradation caused by ASR errors is consistent across LLMs with different absolute performance, suggesting that cascade degradation largely tracks ASR-stage information loss. We further identify single-character Korean ASR errors as a Korean-specific loss channel, where even a minimal transcription difference can change the intended question and degrade downstream QA performance. Finally, an auxiliary comparison shows that a large audio language model outperforms an ASR-LLM cascade with an approximately matched language backbone in noisy Korean SQA, indicating the potential of direct audio input to mitigate transcript-induced information loss.

Synthetic persona datasets cite alignment with official demographics as a basis for trust, yet downstream users consume them as joint structures across age, sex, region, occupation, education, name, and institutional status. Marginal alignment does not imply that these joints are preserved. We propose the Independence-Assumption Footprint (IAF), an audit primitive that operates on the attribute combinations a dataset card itself documents as treated independently. For each such combination, IAF compares the synthetic joint against an external official or institutional reference, using direct joint tables where available and rule-implied checks otherwise. Applied to NVIDIA Nemotron-Personas-Korea (one million Korean synthetic personas), IAF finds that NPK aligns with KOSIS marginals while three joints fail. The major-by-occupation distribution against the KEIS graduate universe carries a large conditional mismatch. The age profile of military service is institutionally inconsistent. Female representation in male-dominated occupations is substantially over-flattened toward parity, with the strict screening verdict mapping-dependent and age-robust under direct standardisation. A transferability demonstration across six further NPK locales finds locale-dependent rather than universal diagnostics, with reference-taxonomy cardinality confounding cross-locale flag counts. For synthetic personas used as silicon samples, marginal claims must therefore be paired with disclosure-anchored joint audits before reuse. The released audit artefacts (reference manifests, occupational crosswalks, derived metrics, reproducibility scripts) instantiate this protocol on the NPK family and are released for retargeting at other synthetic persona resources.

Heterogeneous Graph Prompt Learning (HGPL)has emerged as a promising paradigm for bridging the gap between the objectives of pre-training foundation models and their downstream applications in heterogeneous graph settings. However, existing HGPL methods are primarily designed for in-domain scenarios, whereas real-world deployments often span multiple domains, and the data used for pre-training and downstream tasks may originate from different distributions. Consequently, the applicability of current HGPL approaches is limited to in-domain settings, and their performance typically degrades when application domains shift. To address this serious limitation, we develop CHoE, a cross-domain HGPL method built upon an expert network. During pre-training, we introduce and train structure-conditioned experts, and during prompt tuning, we adopt a structure-aware expert routing and load balancing mechanism to select structurally compatible experts for each meta-path view. In addition, we design a prompt-based semantic fusion module to integrate representations across multiple views for downstream prediction. Extensive experiments show that CHoE consistently improves performance in few-shot cross-domain applications, outperforming all baseline approaches.

Safety evaluations for large language models (LLMs) increasingly target high-stakes National Security and Public Safety (NSPS) risks, yet multilingual safety is mostly assessed through translation-only benchmarks that preserve the underlying scenario, leaving how language and geopolitical context interact largely unexamined beyond a few language pairs. We introduce ROK-FORTRESS, a bilingual, culturally adversarial NSPS benchmark that uses the English-Korean language pair and U.S.-ROK geopolitical axis as a case study, separating the effects of language and geopolitical grounding via a transcreation matrix: adversarial intents are evaluated under controlled combinations of (i) English versus Korean language and (ii) U.S. versus Korean entities, institutions, and operational details. Each adversarial prompt is paired with a dual-use benign counterpart to quantify over-refusal, and responses are scored by calibrated LLM-as-a-judge panels using expert-crafted, prompt-specific binary rubrics. Across a dual-track set of frontier and Korean-optimized models, we find a consistent suppression effect in Korean variants and substantial model-to-model variation in how geopolitical grounding interacts with language; in a subset of models, Korean grounding further mitigates the language-driven suppression. This indicates that, at least in the English-Korean case, safety behavior is shaped by language-as-risk signals and context interactions that translation-only evaluations miss. A direct-request ablation that strips jailbreak wrappers separates a small but persistent reduction for closed-source models from a larger, wrapper-dependent effect that reverses for open-source models, suggesting part of the Korean suppression reflects prompt specialization rather than intrinsic language-based safety alignment. The transcreation matrix methodology is designed to generalize to other language-culture pairs.

This paper aims to construct a linguistic resource of Korean Multiword Expressions for Feature-Based Sentiment Analysis (FBSA): DECO-MWE. Dealing with multiword expressions (MWEs) has been a critical issue in FBSA since many constructs reveal lexical idiosyncrasy. To construct linguistic resources of sentiment MWEs efficiently, we utilize the Local Grammar Graph (LGG) methodology: DECO-MWE is formalized as a Finite-State Transducer that represents lexical-syntactic restrictions on MWEs. In this study, we built a corpus of cosmetics review texts, which show particularly frequent occurrences of MWEs. Based on an empirical examination of the corpus, four types of MWEs have been distinguished. The DECO-MWE thus covers the following four categories: Standard Polarity MWEs (SMWEs), Domain-Dependent Polarity MWEs (DMWEs), Compound Named Entity MWEs (EMWEs) and Compound Feature MWEs (FMWEs). The retrieval performance of the DECO-MWE shows 0.806 f-measure in the test corpus. This study brings a twofold outcome: first, a sizeable general-purpose polarity MWE lexicon, which may be broadly used in FBSA; second, a finite-state methodology adopted in this study to treat domain-dependent MWEs such as idiosyncratic polarity expressions, named entity expressions or feature expressions, and which may be reused in describing linguistic properties of other corpus domains.

Natural language understanding (NLU) is integral to task-oriented dialog systems, but demands a considerable amount of annotated training data to increase the coverage of diverse utterances. In this study, we report the construction of a linguistic resource named FIAD (Financial Annotated Dataset) and its use to generate a Korean annotated training data for NLU in the banking customer service (CS) domain. By an empirical examination of a corpus of banking app reviews, we identified three linguistic patterns occurring in Korean request utterances: TOPIC (ENTITY, FEATURE), EVENT, and DISCOURSE MARKER. We represented them in LGGs (Local Grammar Graphs) to generate annotated data covering diverse intents and entities. To assess the practicality of the resource, we evaluate the performances of DIET-only (Intent: 0.91 /Topic [entity+feature]: 0.83), DIET+ HANBERT (I:0.94/T:0.85), DIET+ KoBERT (I:0.94/T:0.86), and DIET+ KorBERT (I:0.95/T:0.84) models trained on FIAD-generated data to extract various types of semantic items.

Atmospheric visibility is a critical variable for transportation safety and air quality management, however, accurate prediction remains challenging due to the complex interactions between meteorological conditions and air pollutants, as well as the rarity of low-visibility events. This study introduces a machine learning framework to nowcast visibility in six major South Korean cities. To handle the imbalance in the 2018-2020 training data, we applied the Synthetic Minority Over-sampling Technique with Nominal and Continuous (SMOTENC) and Conditional Tabular Generative Adversarial Network (CTGAN). An ensemble approach combining machine learning and deep learning models was then used and evaluated on a 2021 test dataset. The results revealed a marked decline in predictive performance in the test set compared to the cross-validation phase. This degradation was attributed to a distributional shift between training and testing periods, which was quantitatively confirmed by measuring the Wasserstein distance of the most influential feature identified by SHAP analysis. In general, this study presents a methodology that aims to simultaneously address the dual challenges of data imbalance and temporal distributional shifts, and emphasizes the necessity of accounting for evolving external environmental factors when implementing nowcasting models on time-series data.

We report the construction of a Korean evaluation-annotated corpus, hereafter called 'Evaluation Annotated Dataset (EVAD)', and its use in Aspect-Based Sentiment Analysis (ABSA) extended in order to cover e-commerce reviews containing sentiment and non-sentiment linguistic patterns. The annotation process uses Semi-Automatic Symbolic Propagation (SSP). We built extensive linguistic resources formalized as a Finite-State Transducer (FST) to annotate corpora with detailed ABSA components in the fashion e-commerce domain. The ABSA approach is extended, in order to analyze user opinions more accurately and extract more detailed features of targets, by including aspect values in addition to topics and aspects, and by classifying aspectvalue pairs depending whether values are unary, binary, or multiple. For evaluation, the KoBERT and KcBERT models are trained on the annotated dataset, showing robust performances of F1 0.88 and F1 0.90, respectively, on recognition of aspect-value pairs.

Chatbots are robots that can communicate with humans using text or voice signals. Legal chatbots improve access to justice, since legal representation and legal advice by lawyers come with a high cost that excludes disadvantaged and vulnerable people. However, capturing the diversity of actual user input in datasets for deep-learning dialog systems (chatbots) is a technical challenge. Diversity requires large volumes of data, which must also be labelled in order to classify the user's intent, while the cost of labelling datasets increases with volume. Instead of labelling large volumes of authentic data from users, our approach consists in jointly generating large volumes of utterances and high-quality labels. The generator of labelled datasets is based on language resources that take the form of local grammar graphs (LGG), which capture and generalize the vocabulary and local syntax observed by linguists in text. The LGGs associate labels to the utterances according to a domain-specific classification system. We tested this approach by implementing LIGA, a legal chatbot in Korean. The chatbot answers users' conversational queries on legal situations by providing information on similar legal cases, made publicly available by the Korean government. We generated labelled utterances from the LGGs with the aid of the open-source Unitex platform. This process produced 700 million utterances. We trained a DIET classifier on a dataset made of these utterances, and the trained model reached 91% f1-score performance. We implemented a chatbot called LIGA, which uses the results of the model to select a link to a web page that documents similar legal cases.

We propose a simplified human-in-the-loop workflow for second language (L2) Korean morphosyntactic annotation by leveraging agreement between two domain-adapted parsers. We first evaluate whether parser agreement can serve as a proxy for annotation correctness by comparing it with independent human judgments. The results show strong correspondence between parser and human judgments, supporting the feasibility of semi-automatic L2-Korean UD annotation. Further analysis demonstrates that parser disagreements cluster in linguistically predictable domains such as grammatical-relation distinctions and clause-boundary ambiguity. While many disagreement cases are tractable for iterative model refinement, others reflect deeper representational challenges inherent in parsing and tagging L2-Korean corpora.

Large language models (LLMs) are increasingly integrated into legal workflows. However, existing benchmarks primarily address proxy tasks, such as bar examination performance or classification, which fail to capture the performance and risks inherent in day-to-day judicial processes. To address this, we publicly release TriBench-Ko, a Korean benchmark designed to evaluate potential deployment risks of LLMs within the context of verified judicial task requirements. It covers four core tasks: jurisprudence summarization, precedent retrieval, legal issue extraction, and evidence analysis. It jointly assesses model behavior across multiple deployment risk categories, including inaccuracy (hallucination, omission, statutory misapplication), biases (demographic, overcompliance), inconsistencies (prompt sensitivity, non-determinism), and adjudicative overreach. Each item is structured to systematically assess both task performance and a specific risk type based on real judicial decisions. Our evaluation of a range of contemporary LLMs reveals that many models frequently manifest significant risks, most notably struggling with precedent retrieval and failing to capture critical legal information. We provide a comprehensive diagnosis of these LLMs and pinpoint critical areas where LLM-generated outputs in judicial contexts necessitate rigorous inspection and caution. Our dataset and code are available at https://github.com/holi-lab/TriBench-Ko

LLMs and speech assistants are increasingly used for task-oriented interactions, yet their evaluation often relies on controlled scenarios that fail to capture the variability and complexity of real user requests. Drink ordering, for example, involves diverse named entities, drink types, sizes, customizations, and brand-specific terminology, as well as spontaneous speech phenomena such as hesitations and self-corrections. To address this gap, we introduce StarDrinks, a test set in English and Korean containing speech utterances features, transcriptions, and annotated slots. Our dataset supports speech-to-slots SLU, transcription-to-slots NLU, and speech-to-transcription ASR evaluation, providing a realistic benchmark for model robustness and generalization in a linguistically rich, real-world task.

In recent years, the rapid proliferation of open-source large language models (LLMs) has spurred efforts to turn general-purpose models into domain specialists. However, many domain-specialized LLMs are developed using datasets and training protocols that are not aligned with the nuanced requirements of real-world applications. In the legal domain, where precision and reliability are essential, this lack of consideration limits practical utility. In this study, we propose a systematic training framework grounded in the practical needs of the legal domain, with a focus on Korean law. We introduce LegalMidm, a Korean legal-domain LLM, and present a methodology for constructing high-quality, use-case-driven legal datasets and optimized training pipelines. Our approach emphasizes collaboration with legal professionals and rigorous data curation to ensure relevance and factual accuracy, and demonstrates effectiveness in key legal tasks.

Korean aegyo is a socially recognized childlike speaking style used predominantly in romantic interactions among adults. This study examined vowel space modification in aegyo by analyzing formant frequencies from twelve Seoul Korean speakers who produced identical scripts in aegyo and non-aegyo styles. Results show that aegyo speech features a significant increase in F1 values across vowels and selective fronting of front vowels, leading to vowel space expansion but mainly a shift to higher F1. These findings suggest that adult speakers stylize childlike speech by imitating the shorter vocal tract of children, mainly through global vowel lowering and partial fronting.

The development of practical (multimodal) large language model assistants for Korean weather forecasters is hindered by the absence of a multidimensional, expert-level evaluation framework grounded in authoritative sources. To address this, we introduce K-MetBench, a diagnostic benchmark grounded in national qualification exams. It exposes critical gaps across four dimensions: expert visual reasoning of charts, logical validity via expert-verified rationales, Korean-specific geo-cultural comprehension, and fine-grained domain analysis. Our evaluation of 55 models reveals a profound modality gap in interpreting specialized diagrams and a reasoning gap where models hallucinate logic despite correct predictions. Crucially, Korean models outperform significantly larger global models in local contexts, demonstrating that parameter scaling alone cannot resolve cultural dependencies. K-MetBench serves as a roadmap for developing reliable, culturally aware expert AI agents. The dataset is available at https://huggingface.co/datasets/soyeonbot/K-MetBench .

The Korean writing system, \textit{Hangeul}, has a unique character representation rigidly following the invention principles recorded in \textit{Hunminjeongeum}.\footnote{\textit{Hunminjeongeum} is a book published in 1446 that describes the principles of invention and usage of \textit{Hangeul}, devised by King Sejong \cite{Hunminjeongeum_Guide}.} However, existing pre-trained language models (PLMs) for Korean have overlooked these principles. In this paper, we introduce a novel framework for Korean PLMs called KOMBO, which firstly brings the invention principles of \textit{Hangeul} to represent character. Our proposed method, KOMBO, exhibits notable experimental proficiency across diverse NLP tasks. In particular, our method outperforms the state-of-the-art Korean PLM by an average of 2.11% in five Korean natural language understanding tasks. Furthermore, extensive experiments demonstrate that our proposed method is suitable for comprehending the linguistic features of the Korean language. Consequently, we shed light on the superiority of using subcharacters over the typical subword-based approach for Korean PLMs. Our code is available at: https://github.com/SungHo3268/KOMBO.

The intonational structure of Seoul Korean has been defined with discrete tonal categories within the Autosegmental-Metrical model of intonational phonology. However, it is challenging to map continuous $F_0$ contours to these invariant categories due to variable $F_0$ realizations in real-world speech. Our paper proposes Dual-Glob, a deep supervised contrastive learning framework to robustly classify fine-grained pitch accent patterns in Seoul Korean. Unlike conventional local predictive models, our approach captures holistic $F_0$ contour shapes by enforcing structural consistency between clean and augmented views in a shared latent space. To this aim, we introduce the first large-scale benchmark dataset, consisting of manually annotated 10,093 Accentual Phrases in Seoul Korean. Experimental results show that our Dual-Glob significantly outperforms strong baseline models with state-of-the-art accuracy (77.75%) and F1-score (51.54%). Therefore, our work supports AM-based intonational phonology using data-driven methodology, showing that deep contrastive learning effectively captures holistic structural features of continuous $F_0$ contours.

This paper presents a systematic benchmark of state-of-the-art multilingual large language models (LLMs) adapted via token pruning - a compression technique that eliminates tokens and embedding parameters corresponding to languages irrelevant to the target application. Focusing on Korean-centric natural language processing (NLP) tasks, we evaluate architectures including Qwen3, Gemma-3, Llama-3, and Aya across three vocabulary configurations: Original, English-Korean (EnKo), and English-Korean-Chinese (EnKoZh). Performance is assessed using established benchmarks for general aptitude, cultural literacy, instruction following, and machine translation. Our findings indicate that token pruning significantly improves generation stability by eliminating language confusion, and in the case of machine translation, frequently enhances performance on Korean-specific tasks. While instruction-following capabilities display architecture-dependent variance linked to latent cross-lingual representations, the significant reduction in vocabulary size validates token pruning as a highly effective optimization strategy for memory-constrained, domain-specific deployments, despite modest gains in inference latency.

Recent advances in large audio language models (LALMs) have enabled multilingual speech understanding. However, benchmarks for evaluating LALMs remain scarce for non-English languages, with Korean being one such underexplored case. In this paper, we introduce KoALa-Bench, a comprehensive benchmark for evaluating Korean speech understanding and speech faithfulness of LALMs. In particular, KoALa-Bench comprises six tasks. Four tasks evaluate fundamental speech understanding capabilities, including automatic speech recognition, speech translation, speech question answering, and speech instruction following, while the remaining two tasks evaluate speech faithfulness, motivated by our observation that several LALMs often fail to fully leverage the speech modality. Furthermore, to reflect Korea-specific knowledge, our benchmark incorporates listening questions from the Korean college scholastic ability test as well as content covering Korean cultural domains. We conduct extensive experiments across six models, including both white-box and black-box ones. Our benchmark, evaluation code, and leaderboard are publicly available at https://ksbench.github.io/Korean-Benchmark/.

Although negation is known to challenge large language models (LLMs), benchmarks for evaluating negation understanding-especially in Korean-are scarce. We conduct a corpus-based analysis of Korean negation and show that LLM performance degrades under negation. We then introduce Thunder-KoNUBench, a sentence-level negation understanding benchmark that reflects the empirical distribution of Korean negation phenomena. Evaluating 47 LLMs on Thunder-KoNUBench, we analyze the effects of model size and instruction tuning, and perform error analysis to better understand model behavior. We further show that fine-tuning on Thunder-KoNUBench improves negation understanding and broader contextual comprehension in Korean.

The design of Korean constituency treebanks raises a central representational question concerning the choice of terminal units. Although Korean words are morphologically complex, treating morphemes as constituency terminals can obscure the distinction between word-internal morphology and phrase-level syntactic structure, and can create mismatches with eojeol-based dependency resources. This paper argues for an eojeol-based constituency representation, with morphological segmentation and fine-grained POS information encoded in a separate, non-constituent layer. A comparative analysis shows that, under explicit normalization assumptions, the Sejong, Penn Korean, and KAIST treebanks can be compared over a shared eojeol-based constituency backbone. Building on this result, we outline an eojeol-based annotation scheme that preserves interpretable constituency, supports cross-treebank comparison and constituency-dependency alignment, and provides a surface-form terminal layer for future end-to-end Korean constituency parsing.

Accurate and reliable aircraft landing time prediction is essential for effective resource allocation in air traffic management. However, the inherent uncertainty of aircraft trajectories and traffic flows poses significant challenges to both prediction accuracy and trustworthiness. Therefore, prediction models should not only provide point estimates of aircraft landing times but also the uncertainties associated with these predictions. Furthermore, aircraft trajectories are frequently influenced by the presence of nearby aircraft through air traffic control interventions such as radar vectoring. Consequently, landing time prediction models must account for multi-agent interactions in the airspace. In this work, we propose a probabilistic multi-agent aircraft landing time prediction framework that provides the landing times of multiple aircraft as distributions. We evaluate the proposed framework using an air traffic surveillance dataset collected from the terminal airspace of the Incheon International Airport in South Korea. The results demonstrate that the proposed model achieves higher prediction accuracy than the baselines and quantifies the associated uncertainties of its outcomes. In addition, the model uncovered underlying patterns in air traffic control through its attention scores, thereby enhancing explainability.

We present $\textbf{Korean SimpleQA (KoSimpleQA)}$, a benchmark for evaluating factuality in large language models (LLMs) with a focus on Korean cultural knowledge. KoSimpleQA is designed to be challenging yet easy to grade, consisting of 938 short, fact-seeking questions with unambiguous answers. We conduct a comprehensive evaluation across a diverse set of open-source LLMs of varying sizes that support Korean, and find that even the strongest model generates correct answer only 31.6% of the time, underscoring the challenging nature of KoSimpleQA. Notably, performance rankings on KoSimpleQA differ substantially from those on the English SimpleQA, highlighting the unique value of our dataset. Furthermore, we observe that reasoning helps mitigate the cross-lingual knowledge gap in LLMs, which refers to disparities in their ability to manifest knowledge across languages. KoSimpleQA can be found at https://github.com/naver-ai/KoSimpleQA.