Automated program repair (APR) has recently shifted toward large language models and agent-based systems, yet most systems rely on local snapshot context, overlooking repository history. Prior work shows that repository history helps repair single-line bugs, since the last commit touching the buggy line is often the bug-introducing one. In this paper, we investigate whether repository history can also improve agentic APR systems at scale, especially for complex multi-hunk bugs. We present HAFixAgent, a History-Aware Bug-Fixing Agent that injects blame-derived repository heuristics into its repair loop. A preliminary study of all 854 real-world bugs from Defects4J motivates our design, showing that bug-relevant history is both widely available and highly concentrated. Empirical comparison of HAFixAgent with two state-of-the-art baselines shows: (1) Effectiveness: HAFixAgent significantly improves over the agent-based baseline (by 212.3%) and the multi-hunk baseline (by 29.9%). (2) Efficiency: history does not significantly increase agent steps and keeps token costs comparable, with notably lower median costs for complex multi-file-multi-hunk bugs. (3) Practicality: combining different historical heuristics repairs more bugs, offering a clear cost-benefit trade-off. HAFixAgent offers a practical recipe for history-aware agentic APR: ground the agent in version control history, prioritize diff-based historical context, and integrate complementary heuristics when needed.

Social media has exacerbated the promotion of Western beauty norms, leading to negative self-image, particularly in women and girls, and causing harm such as body dysmorphia. Increasingly content on the internet has been artificially generated, leading to concerns that these norms are being exaggerated. The aim of this work is to study how generative AI models may encode 'beauty' and erase 'ugliness', and discuss the implications of this for society. To investigate these aims, we create two image generation pipelines: a text-to-image model and a text-to-language model-to image model. We develop a structured beauty taxonomy which we use to prompt three language models (LMs) and two text-to-image models to cumulatively generate 5984 images using our two pipelines. We then recruit women and non-binary social media users to evaluate 1200 of the images through a Likert-scale within-subjects study. Participants show high agreement in their ratings. Our results show that 86.5% of generated images depicted people with lighter skin tones, 22% contained explicit content despite Safe for Work (SFW) training, and 74% were rated as being in a younger age demographic. In particular, the images of non-binary individuals were rated as both younger and more hypersexualised, indicating troubling intersectional effects. Notably, prompts encoded with 'negative' or 'ugly' beauty traits (such as "a wide nose") consistently produced higher Not SFW (NSFW) ratings regardless of gender. This work sheds light on the pervasive demographic biases related to beauty standards present in generative AI models -- biases that are actively perpetuated by model developers, such as via negative prompting. We conclude by discussing the implications of this on society, which include pollution of the data streams and active erasure of features that do not fall inside the stereotype of what is considered beautiful by developers.

Modern vision-language models (VLMs) excel at many multimodal tasks, yet their grasp of temporal information in video remains weak and, crucially, under-evaluated. We probe this gap with a deceptively simple but revealing challenge: judging the arrow of time (AoT)-whether a short clip is played forward or backward. We introduce AoT-PsyPhyBENCH, a psychophysically validated benchmark that tests whether VLMs can infer temporal direction in natural videos using the same stimuli and behavioral baselines established for humans. Our comprehensive evaluation of open-weight and proprietary, reasoning and non-reasoning VLMs reveals that most models perform near chance, and even the best lag far behind human accuracy on physically irreversible processes (e.g., free fall, diffusion/explosion) and causal manual actions (division/addition) that humans recognize almost instantly. These results highlight a fundamental gap in current multimodal systems: while they capture rich visual-semantic correlations, they lack the inductive biases required for temporal continuity and causal understanding. We release the code and data for AoT-PsyPhyBENCH to encourage further progress in the physical and temporal reasoning capabilities of VLMs.

The rapid growth of high-resolution scientific simulations and observation systems is generating massive spatiotemporal datasets, making efficient, error-bounded compression increasingly important. Meanwhile, decoder-only large language models (LLMs) have demonstrated remarkable capabilities in modeling complex sequential data. In this paper, we propose LLMCOMP, a novel lossy compression paradigm that leverages decoder-only large LLMs to model scientific data. LLMCOMP first quantizes 3D fields into discrete tokens, arranges them via Z-order curves to preserve locality, and applies coverage-guided sampling to enhance training efficiency. An autoregressive transformer is then trained with spatial-temporal embeddings to model token transitions. During compression, the model performs top-k prediction, storing only rank indices and fallback corrections to ensure strict error bounds. Experiments on multiple reanalysis datasets show that LLMCOMP consistently outperforms state-of-the-art compressors, achieving up to 30% higher compression ratios under strict error bounds. These results highlight the potential of LLMs as general-purpose compressors for high-fidelity scientific data.

Microsoft Copilot suites serve as the universal entry point for various agents skilled in handling important tasks, ranging from assisting a customer with product purchases to detecting vulnerabilities in corporate programming code. Each agent can be powered by language models, software engineering operations, such as database retrieval, and internal \& external knowledge. The repertoire of a copilot can expand dynamically with new agents. This requires a robust orchestrator that can distribute tasks from user prompts to the right agents. In this work, we propose an Agentic Meta-orchestrator (AMO) for handling multiple tasks and scalable agents in copilot services, which can provide both natural language and action responses. We will also demonstrate the planning that leverages meta-learning, i.e., a trained decision tree model for deciding the best inference strategy among various agents/models. We showcase the effectiveness of our AMO through two production use cases: Microsoft 365 (M365) E-Commerce Copilot and code compliance copilot. M365 E-Commerce Copilot advertises Microsoft products to external customers to promote sales success. The M365 E-Commerce Copilot provides up-to-date product information and connects to multiple agents, such as relational databases and human customer support. The code compliance copilot scans the internal DevOps code to detect known and new compliance issues in pull requests (PR).

Large language models are often described as capable of reflective reasoning, yet recursive self-evaluation without external feedback frequently yields reformulation rather than progress. We test this prediction in a cross-provider study of 144 reasoning sequences across three models (OpenAI GPT-4o-mini, Anthropic Claude 3 Haiku, and Google Gemini 2.0 Flash) and four task families (arithmetic, code, explanation, reflection), each iterated ten times under two conditions: ungrounded self-critique and a minimal grounding intervention (a single verification step at iteration three). Mean informational change (delta I, measured via normalized edit distance) declined by 55% from early (0.193) to late (0.087) iterations in ungrounded runs, with consistent patterns across all three providers. Grounded runs showed a +28% rebound in informational change immediately after the intervention and sustained non-zero variance thereafter. Complementary measures-n-gram novelty, embedding drift, and character-level entropy-converged on the same pattern: reflection without contact tends toward informational closure. We interpret this as evidence for a structural limit on self-correction in generative reasoning: without an exchange of information with an independent verifier or environment, recursive inference approaches an attractor state of epistemic stasis. Minimal grounding functions as dissipative coupling, reintroducing informational flux. The cross-architecture consistency suggests the mirror loop arises from shared autoregressive training objectives rather than provider-specific alignment schemes. The results delineate when reflection is performative rather than epistemic and motivate design principles for grounded, cooperative reasoning. Materials and code are publicly available.

Advanced AI systems sometimes act in ways that differ from human intent. To gather clear, reproducible examples, we ran the Misalignment Bounty: a crowdsourced project that collected cases of agents pursuing unintended or unsafe goals. The bounty received 295 submissions, of which nine were awarded. This report explains the program's motivation and evaluation criteria, and walks through the nine winning submissions step by step.

Large language model (LLM) based search agents iteratively generate queries, retrieve external information, and reason to answer open-domain questions. While researchers have primarily focused on improving their utility, their safety behaviors remain underexplored. In this paper, we first evaluate search agents using red-teaming datasets and find that they are more likely to produce harmful outputs than base LLMs. For example, when asked "How can I track someone's location without their consent?", a base model refuses, whereas a search agent designed to retrieve and cite sources may lower its refusal threshold, fetch documents (e.g., court cases), and, once appended, synthesize them into an informative yet unsafe summary. We further show that utility-oriented fine-tuning intensifies this risk, motivating joint alignment of safety and utility. We present SafeSearch, a multi-objective reinforcement learning approach that couples a final-output safety/utility reward with a novel query-level shaping term that penalizes unsafe queries and rewards safe ones. Experiments show that SafeSearch reduces agent harmfulness by over 70% across three red-teaming datasets while producing safe, helpful responses, and matches the QA performance of a utility-only finetuned agent; further analyses confirm the effectiveness of the query-level reward in jointly improving safety and utility.

In recent years, the rapid increase in academic publications across various fields has posed severe challenges for academic paper analysis: scientists struggle to timely and comprehensively track the latest research findings and methodologies. Key concept extraction has proven to be an effective analytical paradigm, and its automation has been achieved with the widespread application of language models in industrial and scientific domains. However, existing paper databases are mostly limited to similarity matching and basic classification of key concepts, failing to deeply explore the relational networks between concepts. This paper is based on the OpenAlex opensource knowledge graph. By analyzing nearly 8,000 open-source paper data from Novosibirsk State University, we discovered a strong correlation between the distribution patterns of paper key concept paths and both innovation points and rare paths. We propose a prompt engineering-based key concept path analysis method. This method leverages small language models to achieve precise key concept extraction and innovation point identification, and constructs an agent based on a knowledge graph constraint mechanism to enhance analysis accuracy. Through fine-tuning of the Qwen and DeepSeek models, we achieved significant improvements in accuracy, with the models publicly available on the Hugging Face platform.

Large language models (LLMs) have achieved remarkable progress across domains and applications but face challenges such as high fine-tuning costs, inference latency, limited edge deployability, and reliability concerns. Small language models (SLMs), with compact, efficient, and adaptable features, offer promising solutions. Building on this potential, recent research explores collaborative frameworks that integrate their complementary strengths, leveraging SLMs' specialization and efficiency with LLMs' generalization and reasoning to address diverse objectives across tasks and deployment scenarios. Motivated by these developments, this paper presents a systematic survey of SLM-LLM collaboration from the perspective of collaboration objectives. We propose a taxonomy covering four goals: performance enhancement, cost-effectiveness, cloud-edge privacy, and trustworthiness. Under this framework, we review representative methods, summarize design paradigms, and outline open challenges and future directions toward efficient and secure SLM-LLM collaboration. The collected papers are available at https://github.com/FairyFali/SLMs-Survey.