Large Language Models (LLMs) have demonstrated promising capabilities to generate responses that simulate consistent personality traits. Despite the major attempts to analyze personality expression through output-based evaluations, little is known about how such traits are internally encoded within LLM parameters. In this paper, we introduce a layer-wise probing framework to systematically investigate the layer-wise capability of LLMs in simulating personality for responding. We conduct probing experiments on 11 open-source LLMs over the PersonalityEdit benchmark and find that LLMs predominantly simulate personality for responding in their middle and upper layers, with instruction-tuned models demonstrating a slightly clearer separation of personality traits. Furthermore, by interpreting the trained probing hyperplane as a layer-wise boundary for each personality category, we propose a layer-wise perturbation method to edit the personality expressed by LLMs during inference. Our results show that even when the prompt explicitly specifies a particular personality, our method can still successfully alter the response personality of LLMs. Interestingly, the difficulty of converting between certain personality traits varies substantially, which aligns with the representational distances in our probing experiments. Finally, we conduct a comprehensive MMLU benchmark evaluation and time overhead analysis, demonstrating that our proposed personality editing method incurs only minimal degradation in general capabilities while maintaining low training costs and acceptable inference latency. Our code is publicly available at https://github.com/universe-sky/probing-then-editing-personality.

Methods for navigation based on large-scale learning typically treat each episode as a new problem, where the agent is spawned with a clean memory in an unknown environment. While these generalization capabilities to an unknown environment are extremely important, we claim that, in a realistic setting, an agent should have the capacity of exploiting information collected during earlier robot operations. We address this by introducing a new retrieval-augmented agent, trained with RL, capable of querying a database collected from previous episodes in the same environment and learning how to integrate this additional context information. We introduce a unique agent architecture for the general navigation task, evaluated on ImageNav, Instance-ImageNav and ObjectNav. Our retrieval and context encoding methods are data-driven and employ vision foundation models (FM) for both semantic and geometric understanding. We propose new benchmarks for these settings and we show that retrieval allows zero-shot transfer across tasks and environments while significantly improving performance.

Large Language Models (LLMs) are frequently utilized as sources of knowledge for question-answering. While it is known that LLMs may lack access to real-time data or newer data produced after the model's cutoff date, it is less clear how their knowledge spans across historical information. In this study, we assess the breadth of LLMs' knowledge using financial data of U.S. publicly traded companies by evaluating more than 197k questions and comparing model responses to factual data. We further explore the impact of company characteristics, such as size, retail investment, institutional attention, and readability of financial filings, on the accuracy of knowledge represented in LLMs. Our results reveal that LLMs are less informed about past financial performance, but they display a stronger awareness of larger companies and more recent information. Interestingly, at the same time, our analysis also reveals that LLMs are more likely to hallucinate for larger companies, especially for data from more recent years. The code, prompts, and model outputs are available on GitHub.

In this work, we explore the structure of the embedding space of a transformer model trained for playing a particular reinforcement learning (RL) game. Specifically, we investigate how a transformer-based Proximal Policy Optimization (PPO) model embeds visual inputs in a simple environment where an agent must collect "coins" while avoiding dynamic obstacles consisting of "spotlights. " By adapting Robinson et al. 's [15] study of the volume growth transform for LLMs to the RL setting, we find that the token embedding space for our visual coin collecting game is also not a manifold, and is better modeled as a stratified space, where local dimension can vary from point to point. We further strengthen Robinson's method by proving that fairly general volume growth curves can be realized by stratified spaces. Finally, we carry out an analysis that suggests that as an RL agent acts, its latent representation alternates between periods of low local dimension, while following a fixed sub-strategy, and bursts of high local dimension, where the agent achieves a sub-goal (e.g., collecting an object) or where the environmental complexity increases (e.g., more obstacles appear). Consequently, our work suggests that the distribution of dimensions in a stratified latent space may provide a new geometric indicator of complexity for RL games. 1

Root Cause Analysis (RCA) in mobile networks remains a challenging task due to the need for interpretability, domain expertise, and causal reasoning. In this work, we propose a lightweight framework that leverages Large Language Models (LLMs) for RCA. To do so, we introduce TeleLogs, a curated dataset of annotated troubleshooting problems designed to benchmark RCA capabilities. Our evaluation reveals that existing open-source reasoning LLMs struggle with these problems, underscoring the need for domain-specific adaptation. To address this issue, we propose a two-stage training methodology that combines supervised fine-tuning with reinforcement learning to improve the accuracy and reasoning quality of LLMs. The proposed approach fine-tunes a series of RCA models to integrate domain knowledge and generate structured, multi-step diagnostic explanations, improving both interpretability and effectiveness. Extensive experiments across multiple LLM sizes show significant performance gains over state-of-the-art reasoning and non-reasoning models, including strong generalization to randomized test variants. These results demonstrate the promise of domain-adapted, reasoning-enhanced LLMs for practical and explainable RCA in network operation and management.

In cross-cultural recipe adaptation, the goal is not only to ensure cultural appropriateness and retain the original dish's essence, but also to provide diverse options for various dietary needs and preferences. Retrieval Augmented Generation (RAG) is a promising approach, combining the retrieval of real recipes from the target cuisine for cultural adaptability with large language models (LLMs) for relevance. However, it remains unclear whether RAG can generate diverse adaptation results. Our analysis shows that RAG tends to overly rely on a limited portion of the context across generations, failing to produce diverse outputs even when provided with varied contextual inputs. This reveals a key limitation of RAG in creative tasks with multiple valid answers: it fails to leverage contextual diversity for generating varied responses. To address this issue, we propose CARRIAGE, a plug-and-play RAG framework for cross-cultural recipe adaptation that enhances diversity in both retrieval and context organization. To our knowledge, this is the first RAG framework that explicitly aims to generate highly diverse outputs to accommodate multiple user preferences. Our experiments show that CARRIAGE achieves Pareto efficiency in terms of diversity and quality of recipe adaptation compared to closed-book LLMs.

Millions of individuals' well-being are challenged by the harms of substance use. Harm reduction as a public health strategy is designed to improve their health outcomes and reduce safety risks. Some large language models (LLMs) have demonstrated a decent level of medical knowledge, promising to address the information needs of people who use drugs (PWUD). However, their performance in relevant tasks remains largely unexplored. We introduce HRIPBench, a benchmark designed to evaluate LLM's accuracy and safety risks in harm reduction information provision. The benchmark dataset HRIP-Basic has 2,160 question-answer-evidence pairs. The scope covers three tasks: checking safety boundaries, providing quantitative values, and inferring polysubstance use risks. We build the Instruction and RAG schemes to evaluate model behaviours based on their inherent knowledge and the integration of domain knowledge. Our results indicate that state-of-the-art LLMs still struggle to provide accurate harm reduction information, and sometimes, carry out severe safety risks to PWUD. The use of LLMs in harm reduction contexts should be cautiously constrained to avoid inducing negative health outcomes. WARNING: This paper contains illicit content that potentially induces harms.

Linguistic borrowing is the process of reproducing in one language elements and patterns that come from another language (Hau-gen, 1950). Linguistic borrowing therefore involves the exchange between two languages and has been widely studied within the field of contact linguistics (Weinreich, 1963). Lexical borrowing in particular is the process of importing words from one language into another (Poplack, Sankoff, and Miller, 1988; Onysko, 2007). Lexical borrowing is a phenomenon that occurs in all languages and is a prolific source of new words and meanings (Gerding et al., 2014). In recent decades, English in particular has produced numerous lexical borrowings (often called anglicisms) in many European languages (Furiassi, Pulcini, and Gonz alez, 2012). Previous work estimated that a reader of French newspapers encounters a new lexical borrowing every 1,000 words (Chesley and Baayen, 2010), English borrowings outnumbering all other borrowings combined (Ches-ley, 2010). In Chilean newspapers, lexical borrowings account for approximately 30% of neologisms, 80% of those corresponding to anglicisms (Gerding et al., 2014). In European Spanish, it was estimated that anglicisms could account for 2% of the vocabulary used in Spanish newspaper El Pa ısin 1991 (Rodr ıguez Gonz alez, 2002), a number that is likely to be higher today. As a result, the usage of lexical borrowings in Spanish (and particularly anglicisms) has attracted lots of attention, both in linguistic studies and among the general public.

Machine unlearning aims to remove the influence of specific samples from a trained model. A key challenge in this process is over-unlearning, where the model's performance on the remaining data significantly drops due to the change in the model's parameters. Existing unlearning algorithms depend on the remaining data to prevent this issue. As such, these methods are inapplicable in a more practical scenario, where only the unlearning samples are available (i.e., zero-shot unlearning). This paper presents a novel framework, ZS-P AG, to fill this gap. Our approach offers three key innovations: (1) we approximate the inaccessible remaining data by generating adversarial samples; (2) leveraging the generated samples, we pinpoint a specific subspace to perform the unlearning process, therefore preventing over-unlearning in the challenging zero-shot scenario; and (3) we consider the influence of the unlearning process on the remaining samples and design an influence-based pseudo-labeling strategy. As a result, our method further improves the model's performance after unlearning. The proposed method holds a theoretical guarantee, and experiments on various benchmarks validate the effectiveness and superiority of our proposed method over several baselines.

We introduce a benchmark framework developed by and for the scientific community to evaluate, monitor and steer large language model development in fundamental physics. Building on philosophical concepts of scientific understanding and creativity, we develop a scoring system in which each question is scored by an expert for its correctness, difficulty, and surprise. The questions are of three forms: (i) multiple-choice questions for conceptual understanding, (ii) analytical problems requiring mathematical derivation, and (iii) openended tasks requiring complex problem solving. Our current dataset contains diverse set of examples, including a machine learning challenge to classify high-energy physics events, such as the four top quark signal. To ensure continued relevance, we propose a living benchmark, where physicists contribute questions, for instance alongside new publications. We invite contributions via: http://www.physicsbenchmarks.org/. We hope that this benchmark will enable a targeted AI development that can make a meaningful contribution to fundamental physics research.