Despite the growing demand for multilingual game localization, systematic evaluation frameworks for game translation constraints remain largely unexplored. As part of the WMT26 Test Suite Track, we propose KUEST, a test suite comprising 1,796 English-to-Korean segments across four core categories, and evaluate 27 submitted systems. Experimental results demonstrate that standard metrics obscure true model capabilities (e.g., the overall 1st-place system dropping to 15th in creative translation). Furthermore, we reveal residual capability fragmentation—category-specific variance that survives the shared general-competence factor and inverts top-tier rankings—and observe a "prompt gap” where context metadata degrades performance. As such, KUEST serves not only as a test suite but also as an evaluation framework that multi-dimensionally dissects domain-specific translation blind spots hidden behind single composite scores. The dataset is available at https://huggingface.co/datasets/JudyChoi/KUEST.

Cognitive distortions have been closely linked to mental health disorders, yet their automatic detection remains challenging due to contextual ambiguity, co-occurrence, and semantic overlap. We propose a novel framework that combines Large Language Models (LLMs) with a Multiple-Instance Learning (MIL) architecture to enhance interpretability and expression-level reasoning. Each utterance is decomposed into Emotion, Logic, and Behavior (ELB) components, which are processed by LLMs to infer multiple distortion instances, each with a predicted type, expression, and model-assigned salience score. These instances are integrated via a Multi-View Gated Attention mechanism for final classification. Experiments on Korean (KoACD) and English (Therapist QA) datasets demonstrate that incorporating ELB and LLM-inferred salience scores improves classification performance, especially for distortions with high interpretive ambiguity. Our results suggest a psychologically grounded and generalizable approach for fine-grained reasoning in mental health NLP. The dataset and implementation details are publicly accessible.

Recent research demonstrates that word embeddings, trained on the human-generated corpus, have strong gender biases in embedding spaces, and these biases can result in the discriminative results from the various downstream tasks. Whereas the previous methods project word embeddings into a linear subspace for debiasing, we introduce a Latent Disentanglement method with a siamese auto-encoder structure with an adapted gradient reversal layer. Our structure enables the separation of the semantic latent information and gender latent information of given word into the disjoint latent dimensions. Afterwards, we introduce a Counterfactual Generation to convert the gender information of words, so the original and the modified embeddings can produce a gender-neutralized word embedding after geometric alignment regularization, without loss of semantic information. From the various quantitative and qualitative debiasing experiments, our method shows to be better than existing debiasing methods in debiasing word embeddings. In addition, Our method shows the ability to preserve semantic information during debiasing by minimizing the semantic information losses for extrinsic NLP downstream tasks.