Fine-tuning large language models (LLMs) to adapt to evolving safety policies is costly and impractical. Mechanistic interpretability enables inference-time control through latent activation steering, yet its potential for precise, customizable safety adjustments remains largely untapped. This paper investigates an approach called SafeSteer for guiding the outputs of LLMs by: (i) leveraging category-specific steering vectors for more precise control, (ii) employing a simple, gradient-free unsupervised method to enhance safety steering while preserving text quality, topic relevance, and without explicit refusal, and (iii) accomplishing this without a hard requirement of contrastive pairwise safe data. We also highlight that our method, being simple and effective, aligns with recent studies suggesting that simple techniques often outperform more complex ones in activation steering. We showcase the effectiveness of our approach across various LLMs, datasets, and risk categories, demonstrating its ability to provide precise control, prevent blanket refusals, and guide models toward generating safe content while maintaining topic relevance.

In this short article, I leverage the National Crime Victimization Survey from 1992 to 2022 to examine how income, education, employment, and key demographic factors shape the type of crime victims experience (violent vs property). Using balanced classification splits and logistic regression models evaluated by F1-score, there is an isolation of the socioeconomic drivers of victimization "Group A" models and then an introduction of demographic factors such as age, gender, race, and marital status controls called "Group B" models. The results consistently proves that higher income and education lower the odds of violent relative to property crime, while men younger individuals and racial minorities face disproportionately higher violentcrime risks. On the geographic spectrum, the suburban models achieve the strongest predictive performance with an accuracy of 0.607 and F1 of 0.590, urban areas benefit from adding education and employment predictors and crime in rural areas are still unpredictable using these current factors. The patterns found in this study shows the need for specific interventions like educational investments in metropolitan settings economic support in rural communities and demographicaware prevention strategies.

It has been hypothesized that neural networks with similar architectures trained on similar data learn shared representations relevant to the learning task. We build on this idea by extending the conceptual framework where representations learned across models trained on the same data can be expressed as linear combinations of a \emph{universal} set of basis features. These basis features underlie the learning task itself and remain consistent across models, regardless of scale. From this framework, we propose the \textbf{Linear Representation Transferability (LRT)} Hypothesis -- that there exists an affine transformation between the representation spaces of different models. To test this hypothesis, we learn affine mappings between the hidden states of models of different sizes and evaluate whether steering vectors -- directions in hidden state space associated with specific model behaviors -- retain their semantic effect when transferred from small to large language models using the learned mappings. We find strong empirical evidence that such affine mappings can preserve steering behaviors. These findings suggest that representations learned by small models can be used to guide the behavior of large models, and that the LRT hypothesis may be a promising direction on understanding representation alignment across model scales.

As Large Language Models (LLMs) continue to exhibit increasingly human-like capabilities, aligning them with human values has become critically important. Contemporary advanced techniques, such as prompt learning and reinforcement learning, are being deployed to better align LLMs with human values. However, while these approaches address broad ethical considerations and helpfulness, they rarely focus on simulating individualized human value systems. To address this gap, we present ValueSim, a framework that simulates individual values through the generation of personal backstories reflecting past experiences and demographic information. ValueSim converts structured individual data into narrative backstories and employs a multi-module architecture inspired by the Cognitive-Affective Personality System to simulate individual values based on these narratives. Testing ValueSim on a self-constructed benchmark derived from the World Values Survey demonstrates an improvement in top-1 accuracy by over 10% compared to retrieval-augmented generation methods. Further analysis reveals that performance enhances as additional user interaction history becomes available, indicating the model's ability to refine its persona simulation capabilities over time.

Recent advancements in large language models (LLMs) have revolutionized natural language processing (NLP) and expanded their applications across diverse domains. However, despite their impressive capabilities, LLMs have been shown to reflect and perpetuate harmful societal biases, including those based on ethnicity, gender, and religion. A critical and underexplored issue is the reinforcement of caste-based biases, particularly towards India's marginalized caste groups such as Dalits and Shudras. In this paper, we address this gap by proposing DECASTE, a novel, multi-dimensional framework designed to detect and assess both implicit and explicit caste biases in LLMs. Our approach evaluates caste fairness across four dimensions: socio-cultural, economic, educational, and political, using a range of customized prompting strategies. By benchmarking several state-of-the-art LLMs, we reveal that these models systematically reinforce caste biases, with significant disparities observed in the treatment of oppressed versus dominant caste groups. For example, bias scores are notably elevated when comparing Dalits and Shudras with dominant caste groups, reflecting societal prejudices that persist in model outputs. These results expose the subtle yet pervasive caste biases in LLMs and emphasize the need for more comprehensive and inclusive bias evaluation methodologies that assess the potential risks of deploying such models in real-world contexts.

Multimodal Large Language Models (MLLMs) pose unique safety challenges due to their integration of visual and textual data, thereby introducing new dimensions of potential attacks and complex risk combinations. In this paper, we begin with a detailed analysis aimed at disentangling risks through step-by-step reasoning within multimodal inputs. We find that systematic multimodal risk disentanglement substantially enhances the risk awareness of MLLMs. Via leveraging the strong discriminative abilities of multimodal risk disentanglement, we further introduce \textbf{DREAM} (\textit{\textbf{D}isentangling \textbf{R}isks to \textbf{E}nhance Safety \textbf{A}lignment in \textbf{M}LLMs}), a novel approach that enhances safety alignment in MLLMs through supervised fine-tuning and iterative Reinforcement Learning from AI Feedback (RLAIF). Experimental results show that DREAM significantly boosts safety during both inference and training phases without compromising performance on normal tasks (namely oversafety), achieving a 16.17\% improvement in the SIUO safe\&effective score compared to GPT-4V. The data and code are available at https://github.com/Kizna1ver/DREAM.

--In recent years, Unsupervised Domain Adaptation (UDA) has gained significant attention in the field of Natural Language Processing (NLP) owing to its ability to enhance model generalization across diverse domains. However, its application for knowledge transfer between distinct legal domains remains largely unexplored. T o address the challenges posed by lengthy and complex legal texts and the limited availability of large-scale annotated datasets, we propose JurisCTC, a novel model designed to improve the accuracy of Legal Judgment Prediction (LJP) tasks. Unlike existing approaches, JurisCTC facilitates effective knowledge transfer across various legal domains and employs contrastive learning to distinguish samples from different domains. Specifically, for the LJP task, we enable knowledge transfer between civil and criminal law domains. Compared to other models and specific large language models (LLMs), JurisCTC demonstrates notable advancements, achieving peak accuracies of 76.59% and 78.83%, respectively. Legal Judgment Prediction (LJP) refers to the task of forecasting court outcomes based on the facts of a legal case, as well as other relevant information such as arguments and claims presented in the case description. This field aims to leverage computational techniques to predict judicial decisions, offering significant benefits across various legal contexts. Automated LJP systems have considerable practical value: they can assist legal professionals in analyzing cases and providing consultation services to the public, thereby reducing legal costs and improving access to justice.

Introduction Swarm robotics systems ( 1) have the potential to revolutionize many industries, from targeted material delivery ( 2) to precision farming ( 3, 4). Boosted by technical breakthroughs, such as cloud computing ( 5, 6), novel hardware design ( 7-9), and manufacturing techniques ( 10), swarms of robots are envisioned to play an important role in both industrial ( 12) and urban ( 13, 14) activities. The emergence of robot swarms has been acknowledged as one of the ten robotics grand challenges for the next 5-10 years that will have significant socioeconomic impact. Despite having such a promising future, many important aspects which need to be considered in realistic deployments are either underexplored or neglected ( 15). One of the main reasons why swarms of robots have not been widely adopted in real-world applications is because there is no consensus on how to design swarm robotics systems that include perception, action, and communication ( 15). In addition, recent research points out that the lack of security standards in the field is also hindering the adoption of this technology in data-sensitive areas (e.g., military, surveillance, public infrastructure) ( 16, 17). These research gaps are motivating scientists to focus on new fields of study such as applied swarm security (18, 19) and privacy ( 20, 21) as well as to revisit already accepted assumptions in the field. From the origins of swarm robotics research, robot swarms were assumed to be fault-tolerant by design, due to the large number of robot units involved ( 22-25). However, it has been shown that a small number of partially failed (with defective sensors, broken actuators, noisy communications devices, etc.) ( 26) or malicious robots ( 27,28) can have a significant impact on 2 Figure 1: T owards secure and secret cooperation in swarm robotics missions.

This paper challenges the assumption that courts should grant First Amendment protections to outputs from large generative AI models, such as GPT-4 and Gemini. We argue that because these models lack intentionality, their outputs do not constitute speech as understood in the context of established legal precedent, so there can be no speech to protect. Furthermore, if the model outputs are not speech, users cannot claim a First Amendment speech right to receive the outputs. We also argue that extending First Amendment rights to AI models would not serve the fundamental purposes of free speech, such as promoting a marketplace of ideas, facilitating self-governance, or fostering self-expression. In fact, granting First Amendment protections to AI models would be detrimental to society because it would hinder the government's ability to regulate these powerful technologies effectively, potentially leading to the unchecked spread of misinformation and other harms.

Large language models (LLMs) are typically trained on enormous quantities of unlicensed text, a practice that has led to scrutiny due to possible intellectual property infringement and ethical concerns. Training LLMs on openly licensed text presents a first step towards addressing these issues, but prior data collection efforts have yielded datasets too small or low-quality to produce performant LLMs. To address this gap, we collect, curate, and release the Common Pile v0.1, an eight terabyte collection of openly licensed text designed for LLM pretraining. The Common Pile comprises content from 30 sources that span diverse domains including research papers, code, books, encyclopedias, educational materials, audio transcripts, and more. Crucially, we validate our efforts by training two 7 billion parameter LLMs on text from the Common Pile: Comma v0.1-1T and Comma v0.1-2T, trained on 1 and 2 trillion tokens respectively. Both models attain competitive performance to LLMs trained on unlicensed text with similar computational budgets, such as Llama 1 and 2 7B. In addition to releasing the Common Pile v0.1 itself, we also release the code used in its creation as well as the training mixture and checkpoints for the Comma v0.1 models.