This paper analyzes the multilinguality of LLMs by examining their structural properties, finding that low-resource languages are structurally more distinct from English than high-resource languages,…

The paper identifies a failure mode called spatial lexical bias in MLLMs, where adding a spatial word to options biases the model's choice, and demonstrates that this failure originates primarily from…

The paper audits six LLMs across four languages, finding that their gender stereotyping is significantly wider than human baselines and that cross-lingual translation fundamentally alters the nature o…

The paper demonstrates that the AI-like style introduced by post-training alignment can be measured, localized, and causally removed using a novel ablation technique called PASTA.

The paper introduces BiAxisAudit, a novel framework that evaluates LLM bias by analyzing bias scores across multiple prompt formats and within the internal inconsistency of model responses, revealing…

This study demonstrates that Chain-of-Thought (CoT) monitoring is fundamentally fragile and unreliable for detecting misaligned behavior across typologically diverse languages, especially in low-resou…

The paper challenges the conclusion that LLMs lack reasoning by demonstrating that reported performance drops on GSM-Symbolic are often statistically weak and partially attributable to dataset biases,…

The paper introduces the Triangulated Preference Shift score, an automated, curation-free metric to quantify systematic lexical biases introduced into Large Language Models during the preference-learn…

The paper proposes a zero-shot cross-lingual method to estimate language model confidence by training a lightweight linear probe on one language and applying it directly to unseen, typologically diver…

dMoE proposes a block-level Mixture-of-Experts (MoE) framework for Diffusion Large Language Models (dLLMs) that aggregates token-level expert distributions into a unified block-level distribution, sig…

The paper introduces RA-MoE, a novel fine-tuning framework that leverages the internal routing structure of Mixture-of-Experts (MoE) models to improve performance on multilingual downstream tasks by a…

The paper introduces MLLM-Microscope, a system that analyzes the internal structure of multimodal large language models (MLLMs), finding that modality fusion significantly impacts the linearity and di…

The paper introduces a multilingual benchmark (MentalMap) to test if LLMs build internal spatial world models from text, finding a universal 'L3 reasoning cliff' suggesting that text-only working memo…

The paper introduces XLGoBench, a synthetic benchmark of algorithmic tasks designed to detect persistent cross-lingual skill gaps in large language models.

The paper proposes a debiasing fine-tuning technique to efficiently enhance the robustness of Large Language Models against semantically similar but textually altered prompts.

The paper introduces three new Korean speech benchmarks (KVoiceBench, KOpenAudioBench, and KMMAU) to evaluate SpeechLMs, demonstrating that English-centric evaluation fails to capture performance gaps…

The paper proposes MIMO, a two-stage framework that improves Multilingual Information Retrieval (MLIR) by stabilizing cross-lingual alignment and enhancing retrieval discrimination using a combination…

The paper systematically evaluates concept-based explainability in MLLMs, finding that forcing models to generate formal explanations degrades predictive accuracy, suggesting that explaining is genuin…

This paper systematically analyzes how different architectural components of Large Vision-Language Models (LVLMs) contribute to hallucination robustness, finding that joint enhancement of visual fidel…