Driven by privacy protection laws and regulations, unlearning in Large Language Models (LLMs) is gaining increasing attention. However, current research often neglects the interpretability of the unlearning process, particularly concerning sample-level unlearning difficulty. Existing studies typically assume a uniform unlearning difficulty across samples. This simplification risks attributing the performance of unlearning algorithms to sample selection rather than the algorithm's design, potentially steering the development of LLM unlearning in the wrong direction. Thus, we investigate the relationship between LLM unlearning and sample characteristics, with a focus on unlearning difficulty. Drawing inspiration from neuroscience, we propose a Memory Removal Difficulty ($\mathrm{MRD}$) metric to quantify sample-level unlearning difficulty. Using $\mathrm{MRD}$, we analyze the characteristics of hard-to-unlearn versus easy-to-unlearn samples. Furthermore, we propose an $\mathrm{MRD}$-based weighted sampling method to optimize existing unlearning algorithms, which prioritizes easily forgettable samples, thereby improving unlearning efficiency and effectiveness. We validate the proposed metric and method using public benchmarks and datasets, with results confirming its effectiveness.

Pre-trained large language models (LLMs) have been demonstrated to possess intrinsic reasoning capabilities that can emerge naturally when expanding the response space. However, the neural representation mechanisms underlying these intrinsic capabilities and approaches for their optimal utilization remain inadequately understood. In this work, we make the key discovery that a simple linear classifier can effectively detect intrinsic reasoning capabilities in LLMs' activation space, particularly within specific representation types and network layers. Based on this finding, we propose a classifier-guided search framework that strategically explore a tree-structured response space. In each node expansion, the classifier serves as a scoring and ranking mechanism that efficiently allocates computational resources by identifying and prioritizing more thoughtful reasoning directions for continuation. After completing the tree expansion, we collect answers from all branches to form a candidate answer pool. We propose a branch-aggregation selection method that marginalizes over all supporting branches by aggregating their thoughtfulness scores, thereby identifying the optimal answer from the pool. Experimental results show that our framework's comprehensive exploration not only covers valid reasoning chains but also effectively identifies them, achieving significant improvements across multiple arithmetic reasoning benchmarks.

Business processes are fundamental to organizational operations, yet their optimization remains challenging due to the time-consuming nature of manual process analysis. Our paper harnesses Large Language Models (LLMs) to automate value-added analysis, a qualitative process analysis technique that aims to identify steps in the process that do not deliver value. To date, this technique is predominantly manual, time-consuming, and subjective. Our method offers a more principled approach which operates in two phases: first, decomposing high-level activities into detailed steps to enable granular analysis, and second, performing a value-added analysis to classify each step according to Lean principles. This approach enables systematic identification of waste while maintaining the semantic understanding necessary for qualitative analysis. We develop our approach using 50 business process models, for which we collect and publish manual ground-truth labels. Our evaluation, comparing zero-shot baselines with more structured prompts reveals (a) a consistent benefit of structured prompting and (b) promising performance for both tasks. We discuss the potential for LLMs to augment human expertise in qualitative process analysis while reducing the time and subjectivity inherent in manual approaches.

In this paper, we propose a collaborative shelf-picking framework that combines multimodal interaction, physics-based reasoning, and task division for enhanced human-robot teamwork. The framework enables the robot to recognize human pointing gestures, interpret verbal cues and voice commands, and communicate through visual and auditory feedback. Moreover, it is powered by a Large Language Model (LLM) which utilizes Chain of Thought (CoT) and a physics-based simulation engine for safely retrieving cluttered stacks of boxes on shelves, relationship graph for sub-task generation, extraction sequence planning and decision making.

Imitation learning based planning tasks on the nuPlan dataset have gained great interest due to their potential to generate human-like driving behaviors. However, open-loop training on the nuPlan dataset tends to cause causal confusion during closed-loop testing, and the dataset also presents a long-tail distribution of scenarios. These issues introduce challenges for imitation learning. To tackle these problems, we introduce CAFE-AD, a Cross-Scenario Adaptive Feature Enhancement for Trajectory Planning in Autonomous Driving method, designed to enhance feature representation across various scenario types. We develop an adaptive feature pruning module that ranks feature importance to capture the most relevant information while reducing the interference of noisy information during training. Moreover, we propose a cross-scenario feature interpolation module that enhances scenario information to introduce diversity, enabling the network to alleviate over-fitting in dominant scenarios. We evaluate our method CAFE-AD on the challenging public nuPlan Test14-Hard closed-loop simulation benchmark. The results demonstrate that CAFE-AD outperforms state-of-the-art methods including rule-based and hybrid planners, and exhibits the potential in mitigating the impact of long-tail distribution within the dataset. Additionally, we further validate its effectiveness in real-world environments. The code and models will be made available at https://github.com/AlniyatRui/CAFE-AD.

Large language models (LLMs) offer a promising pre-screening tool, improving early disease detection and providing enhanced healthcare access for underprivileged communities. The early diagnosis of various diseases continues to be a significant challenge in healthcare, primarily due to the nonspecific nature of early symptoms, the shortage of expert medical practitioners, and the need for prolonged clinical evaluations, all of which can delay treatment and adversely affect patient outcomes. With impressive accuracy in prediction across a range of diseases, LLMs have the potential to revolutionize clinical pre-screening and decision-making for various medical conditions. In this work, we study the diagnostic capability of LLMs for Rheumatoid Arthritis (RA) with real world patients data. Patient data was collected alongside diagnoses from medical experts, and the performance of LLMs was evaluated in comparison to expert diagnoses for RA disease prediction. We notice an interesting pattern in disease diagnosis and find an unexpected \textit{misalignment between prediction and explanation}. We conduct a series of multi-round analyses using different LLM agents. The best-performing model accurately predicts rheumatoid arthritis (RA) diseases approximately 95\% of the time. However, when medical experts evaluated the reasoning generated by the model, they found that nearly 68\% of the reasoning was incorrect. This study highlights a clear misalignment between LLMs high prediction accuracy and its flawed reasoning, raising important questions about relying on LLM explanations in clinical settings. \textbf{LLMs provide incorrect reasoning to arrive at the correct answer for RA disease diagnosis.}

In this paper, we show it is possible to bypass the safety guardrails of large language models (LLMs) through a humorous prompt including the unsafe request. In particular, our method does not edit the unsafe request and follows a fixed template -- it is simple to implement and does not need additional LLMs to craft prompts. Extensive experiments show the effectiveness of our method across different LLMs. We also show that both removing and adding more humor to our method can reduce its effectiveness -- excessive humor possibly distracts the LLM from fulfilling its unsafe request. Thus, we argue that LLM jailbreaking occurs when there is a proper balance between focus on the unsafe request and presence of humor.

Large language models (LLMs) have achieved impressive results in a wide range of natural language applications. However, they often struggle to recognize low-resource languages, in particular African languages, which are not well represented in large training corpora. In this paper, we consider how to adapt LLMs to low-resource African languages. We find that combining curated data from African languages with high-quality English educational texts results in a training mix that substantially improves the model's performance on these languages. On the challenging IrokoBench dataset, our models consistently achieve the best performance amongst similarly sized baselines, particularly on knowledge-intensive multiple-choice questions (AfriMMLU). Additionally, on the cross-lingual question answering benchmark AfriQA, our models outperform the base model by over 10%. To better understand the role of English data during training, we translate a subset of 200M tokens into Swahili language and perform an analysis which reveals that the content of these data is primarily responsible for the strong performance. We release our models and data to encourage future research on African languages.

For decades, Moore's Law has served as a steadfast pillar in computer architecture and system design, promoting a clear abstraction between hardware and software. This traditional Moore's computing paradigm has deepened the rift between the two, enabling software developers to achieve near-exponential performance gains often without needing to delve deeply into hardware-specific optimizations. Yet today, Moore's Law -- with its once relentless performance gains now diminished to incremental improvements -- faces inevitable physical barriers. This stagnation necessitates a reevaluation of the conventional system design philosophy. The traditional decoupled system design philosophy, which maintains strict abstractions between hardware and software, is increasingly obsolete. The once-clear boundary between software and hardware is rapidly dissolving, replaced by co-design. It is imperative for the computing community to intensify its commitment to hardware-software co-design, elevating system abstractions to first-class citizens and reimagining design principles to satisfy the insatiable appetite of modern computing. Hardware-software co-design is not a recent innovation. To illustrate its historical evolution, I classify its development into five relatively distinct ``epochs''. This post also highlights the growing influence of the architecture community in interdisciplinary teams -- particularly alongside ML researchers -- and explores why current co-design paradigms are struggling in today's computing landscape. Additionally, I will examine the concept of the ``hardware lottery'' and explore directions to mitigate its constraining influence on the next era of computing innovation.

As telecommunication service providers shifting their focus to analyzing user behavior for package design and marketing interventions, a critical challenge lies in developing a unified, end-to-end framework capable of modeling long-term and periodic user behavior sequences with diverse time granularities, multi-modal data inputs, and heterogeneous labels. This paper introduces GTS-LUM, a novel user behavior model that redefines modeling paradigms in telecommunication settings. GTS-LUM adopts a (multi-modal) encoder-adapter-LLM decoder architecture, enhanced with several telecom-specific innovations. Specifically, the model incorporates an advanced timestamp processing method to handle varying time granularities. It also supports multi-modal data inputs -- including structured tables and behavior co-occurrence graphs -- and aligns these with semantic information extracted by a tokenizer using a Q-former structure. Additionally, GTS-LUM integrates a front-placed target-aware mechanism to highlight historical behaviors most relevant to the target. Extensive experiments on industrial dataset validate the effectiveness of this end-to-end framework and also demonstrate that GTS-LUM outperforms LLM4Rec approaches which are popular in recommendation systems, offering an effective and generalizing solution for user behavior modeling in telecommunications.