Human feedback plays a pivotal role in aligning large language models (LLMs) with human preferences. However, such feedback is often noisy or inconsistent, which can degrade the quality of reward models and hinder alignment. While various automated data cleaning methods have been proposed to mitigate this issue, a systematic evaluation of their effectiveness and generalizability remains lacking. To bridge this gap, we introduce the first comprehensive benchmark for evaluating 13 preference data cleaning methods in the context of LLM alignment. PrefCleanBench offers a standardized protocol to assess cleaning strategies in terms of alignment performance and generalizability across diverse datasets, model architectures, and optimization algorithms. By unifying disparate methods and rigorously comparing them, we uncover key factors that determine the success of data cleaning in alignment tasks. This benchmark lays the groundwork for principled and reproducible approaches to improving LLM alignment through better data quality-highlighting the crucial but underexplored role of data preprocessing in responsible AI development. We release modular implementations of all methods to catalyze further research: https://github.com/deeplearning-wisc/PrefCleanBench.

Large language models (LLMs) excel at complex reasoning but can still exhibit harmful behaviors. Current alignment strategies typically embed safety into model weights, making these controls implicit, static, and difficult to modify. This paper introduces Cognition-of-Thought (CooT), a novel decoding-time framework that equips LLMs with an explicit cognitive self-monitoring loop. CooT couples a standard text Generator with a cognitive Perceiver that continuously monitors the unfolding sequence. The Perceiver uses a structured, precedence-based hierarchy of principles (e.g., safety over obedience) to detect potential misalignments as they arise. When violations are flagged, CooT intervenes by rolling back the generation to the point of error and regenerating under injected guidance that combines universal social priors with context-specific warnings. CooT thus transforms alignment from a fixed property into an explicit, dynamic, and auditable process active during inference, allowing for flexible policy updates without retraining the model. Extensive experiments across multiple benchmarks and model families confirm that CooT consistently improves safety and social reasoning performance.

Retrieval-Augmented Generation (RAG) aims to mitigate hallucinations in large language models (LLMs) by grounding responses in retrieved documents. Yet, RAG-based LLMs still hallucinate even when provided with correct and sufficient context. A growing line of work suggests that this stems from an imbalance between how models use external context and their internal knowledge, and several approaches have attempted to quantify these signals for hallucination detection. However, existing methods require extensive hyperparameter tuning, limiting their generalizability. We propose LUMINA, a novel framework that detects hallucinations in RAG systems through context-knowledge signals: external context utilization is quantified via distributional distance, while internal knowledge utilization is measured by tracking how predicted tokens evolve across transformer layers. We further introduce a framework for statistically validating these measurements. Experiments on common RAG hallucination benchmarks and four open-source LLMs show that LUMINA achieves consistently high AUROC and AUPRC scores, outperforming prior utilization-based methods by up to +13% AUROC on HalluRAG. Moreover, LUMINA remains robust under relaxed assumptions about retrieval quality and model matching, offering both effectiveness and practicality.

KT developed a Responsible AI (RAI) assessment methodology and risk mitigation technologies to ensure the safety and reliability of AI services. By analyzing the Basic Act on AI implementation and global AI governance trends, we established a unique approach for regulatory compliance and systematically identify and manage all potential risk factors from AI development to operation. We present a reliable assessment methodology that systematically verifies model safety and robustness based on KT's AI risk taxonomy tailored to the domestic environment. We also provide practical tools for managing and mitigating identified AI risks. With the release of this report, we also release proprietary Guardrail : SafetyGuard that blocks harmful responses from AI models in real-time, supporting the enhancement of safety in the domestic AI development ecosystem. We also believe these research outcomes provide valuable insights for organizations seeking to develop Responsible AI.

Large language models (LLMs) typically deploy safety mechanisms to prevent harmful content generation. Most current approaches focus narrowly on risks posed by malicious actors, often framing risks as adversarial events and relying on defensive refusals. However, in real-world settings, risks also come from non-malicious users seeking help while under psychological distress (e.g., self-harm intentions). In such cases, the model's response can strongly influence the user's next actions. Simple refusals may lead them to repeat, escalate, or move to unsafe platforms, creating worse outcomes. We introduce Constructive Safety Alignment (CSA), a human-centric paradigm that protects against malicious misuse while actively guiding vulnerable users toward safe and helpful results. Implemented in Oyster-I (Oy1), CSA combines game-theoretic anticipation of user reactions, fine-grained risk boundary discovery, and interpretable reasoning control, turning safety into a trust-building process. Oy1 achieves state-of-the-art safety among open models while retaining high general capabilities. On our Constructive Benchmark, it shows strong constructive engagement, close to GPT-5, and unmatched robustness on the Strata-Sword jailbreak dataset, nearing GPT-o1 levels. By shifting from refusal-first to guidance-first safety, CSA redefines the model-user relationship, aiming for systems that are not just safe, but meaningfully helpful. We release Oy1, code, and the benchmark to support responsible, user-centered AI.

Pedestrians Localization in Non-Line-of-Sight (NLoS) regions within urban environments poses a significant challenge for autonomous driving systems. While mmWave radar has demonstrated potential for detecting objects in such scenarios, the 2D radar point cloud (PCD) data is susceptible to distortions caused by multipath reflections, making accurate spatial inference difficult. Additionally, although camera images provide high-resolution visual information, they lack depth perception and cannot directly observe objects in NLoS regions. In this paper, we propose a novel framework that interprets radar PCD through road layout inferred from camera for localization of NLoS pedestrians. The proposed method leverages visual information from the camera to interpret 2D radar PCD, enabling spatial scene reconstruction. The effectiveness of the proposed approach is validated through experiments conducted using a radar-camera system mounted on a real vehicle. The localization performance is evaluated using a dataset collected in outdoor NLoS driving environments, demonstrating the practical applicability of the method.

OpenAI plans to allow a wider range of content, including erotica, on its popular chatbot ChatGPT as part of its push to treat adult users like adults, says its boss Sam Altman. In a post on X on Tuesday, Mr Altman said upcoming versions of the popular chatbot would enable it to behave in a more human-like way - but only if you want it, not because we are usage maxxing. The move, reminiscent of Elon Musk's xAI recent introduction of two sexually explicit chatbots to Grok, could help OpenAI attract more paying subscribers. It is also likely to intensify pressure on lawmakers to introduce tighter restrictions on chatbot companions. OpenAI did not respond to the BBC's requests for comment following Mr Altman's post.

Another round of terminations, combined with previous layoffs and departures, has reduced the Centers for Disease Control and Prevention workforce by about 3,000 people since January. After the latest round of mass firings at the Centers for Disease Control and Prevention over the weekend, the union that represents agency employees estimates that around 3,000 people this year--about a quarter of the agency's workforce--have departed the agency. That number includes workers affected by layoffs earlier this year, as well those who have accepted the Trump administration's "Fork in the Road" buyout program. The most recent cuts came down amidst the ongoing government shutdown. On October 10, more than 1,300 CDC employees received termination notices.

The company launched a dedicated ChatGPT experience for under-18 users in September. The company launched a dedicated ChatGPT experience for under-18 users in September. New version will allow users to customize AI assistant's personality in what firm calls'treat adults users like adults' policy OpenAI announced plans on Tuesday to relax restrictions on its ChatGPT chatbot, including allowing erotic content for verified adult users as part of what the company calls a "treat adult users like adults" principle. OpenAI's plan includes the release of an updated version of ChatGPT that will allow users to customize their AI assistant's personality, including options for more human-like responses, heavy emoji use, or friend-like behavior. The most significant change will come in December, when OpenAI plans to roll out more comprehensive age-gating that would permit erotic content for adults who have verified their ages.

Many Say They're Not Involved A presentation that has been shared with the Trump administration references Tesla, Ikea, TSMC, and more in its plan to rebuild Gaza. Some of these companies say they had no idea they were mentioned. The mound of rubble at the site of the Unknown Soldier Tower, destroyed by overnight Israeli bombardment, is pictured in the Rimal neighbourhood of Gaza City on September 15, 2025. A sweeping plan to reconstruct Gaza, which has been shared with Trump administration officials, features the names and logos of more than two dozen companies--some of which tell WIRED they had no knowledge they were named or involved. The presentation outlining the plan was reportedly created by some of the businessmen who helped ideate what became the controversial nonprofit the Gaza Humanitarian Foundation, which is currently leading aid distribution in Gaza, calling for the creation of a new entity called the Gaza Reconstitution, Economic Acceleration and Transformation (GREAT) Trust.