Alignment, interpretability, red-teaming, bias and privacy: the research on what these systems do when they misbehave.
arXiv:2608.29109v1 Announce Type: new Abstract: Large language models often answer structurally unanswerable questions, such as computing cot(-540{\deg}) or evaluating (1).startswith("1"), instead of...
arXiv:2608.29431v1 Announce Type: new Abstract: Localized Representation Steering (LRS) is widely used to correct reasoning pathologies in large language models. However, standard benchmark evaluatio...
arXiv:2608.29610v1 Announce Type: new Abstract: The current alignment tuning paradigm for Large Language Models (LLMs) prioritizes surface-level behaviors -- fluency, safety, and tonal consistency. W...
arXiv:2608.29624v1 Announce Type: new Abstract: Natural Language Processing methods have enabled novel solutions and advances in the field of privacy, particularly in the sub-domain of text-to-text p...
arXiv:2608.29948v1 Announce Type: new Abstract: Evaluating data-text alignment remains challenging: existing metrics often provide limited explanations for the scores, while prompt-based LLM-as-Judge...
arXiv:2608.30197v1 Announce Type: new Abstract: Safety alignment is essential for deploying large language models, requiring systems to prevent harmful compliance while preserving helpfulness on beni...
arXiv:2608.30303v1 Announce Type: new Abstract: Search agents reduce hallucination by grounding answers in retrieved web evidence. Yet reliance on retrieval also creates an attack surface: poisoned c...
arXiv:2608.30373v1 Announce Type: new Abstract: Multi-Agent Debate (MAD) has been widely adopted to improve LLM-based evaluation by prompting multiple agents to negotiate and reach a consensus. Howev...
arXiv:2608.30426v1 Announce Type: new Abstract: Current dialogue systems struggle with dynamic information retrieval, often leading to hallucinations and lower response accuracy. We address this by a...
arXiv:2608.31105v1 Announce Type: new Abstract: Users of a deployed language model routinely encounter behaviours that testing almost never surfaces, since deployment puts the model through orders of...
arXiv:2608.30649v1 Announce Type: new Abstract: Removing a specific individual's information from multimodal large language models (MLLMs) is often needed after deployment, but existing methods rely...
arXiv:2608.30754v1 Announce Type: new Abstract: Evaluating creativity in large language model (LLM) outputs remains challenging because creativity is multidimensional and human-centered. We examine h...
arXiv:2608.30842v1 Announce Type: new Abstract: Humans play a vital role at every stage of AI development, from data collection and curation to model development and evaluation. However, humans often...
arXiv:2608.30948v1 Announce Type: new Abstract: LLMs can imitate how people write, which raises concerns about impersonation, trust, and detection in social settings. These concerns are especially im...
arXiv:2608.31035v1 Announce Type: new Abstract: Codec-based text-to-speech (TTS) models make language-model post-training applicable to speech generation, but it remains unclear when learned perceptu...
arXiv:2608.31038v1 Announce Type: new Abstract: Incremental Named Entity Recognition (INER) stands as a pivotal task in information extraction, emphasizing the successive identification of new entity...
arXiv:2608.29483v1 Announce Type: cross Abstract: Modern Vision-Language Models (VLMs) perform well above the human baseline in image geolocalization, a task critically important in disaster response...
arXiv:2608.29847v1 Announce Type: cross Abstract: Text-to-image models learn associations between concepts - in the case of this paper, people's professions, which we refer to as roles - and visual a...
arXiv:2604.07518v2 Announce Type: replace Abstract: Vision-Language Models often struggle with complex visual reasoning due to the visual information loss in textual CoT. Existing methods either add...
arXiv:2604.25580v2 Announce Type: replace Abstract: Perspective API closes at the end of 2026, removing the de facto standard for toxicity measurement and exposing researchers' dependence on a tool t...