The expansion of large language models (LLMs) with hundreds of billions of parameters presents significant challenges to computational resources, particularly data movement and memory bandwidth. Long-context LLMs, which process sequences of tens of thousands of tokens, further increase the demand on the memory system as the complexity in attention layers and key-value cache sizes is proportional to the context length. Processing-in-Memory (PIM) maximizes memory bandwidth by moving compute to the data and can address the memory bandwidth challenges; however, PIM is not necessarily scalable to accelerate long-context LLM because of limited per-module memory capacity and the inflexibility of fixed-functional unit PIM architecture and static memory management. In this work, we propose LoL-PIM which is a multi-node PIM architecture that accelerates long context LLM through hardware-software co-design. In particular, we propose how pipeline parallelism can be exploited across a multi-PIM module while a direct PIM access (DPA) controller (or DMA for PIM) is proposed that enables dynamic PIM memory management and results in efficient PIM utilization across a diverse range of context length. We developed an MLIR-based compiler for LoL-PIM extending a commercial PIM-based compiler where the software modifications were implemented and evaluated, while the hardware changes were modeled in the simulator. Our evaluations demonstrate that LoL-PIM significantly improves throughput and reduces latency for long-context LLM inference, outperforming both multi-GPU and GPU-PIM systems (up to 8.54x and 16.0x speedup, respectively), thereby enabling more efficient deployment of LLMs in real-world applications.

Despite their impressive ability to generate high-quality and fluent text, generative large language models (LLMs) also produce hallucinations: statements that are misaligned with established world knowledge or provided input context. However, measuring hallucination can be challenging, as having humans verify model generations on-the-fly is both expensive and time-consuming. In this work, we release HALoGEN, a comprehensive hallucination benchmark consisting of: (1) 10,923 prompts for generative models spanning nine domains including programming, scientific attribution, and summarization, and (2) automatic high-precision verifiers for each use case that decompose LLM generations into atomic units, and verify each unit against a high-quality knowledge source. We use this framework to evaluate ~150,000 generations from 14 language models, finding that even the best-performing models are riddled with hallucinations (sometimes up to 86% of generated atomic facts depending on the domain). We further define a novel error classification for LLM hallucinations based on whether they likely stem from incorrect recollection of training data (Type A errors), or incorrect knowledge in training data (Type B errors), or are fabrication (Type C errors). We hope our framework provides a foundation to enable the principled study of why generative models hallucinate, and advances the development of trustworthy large language models.

Much of the research focus on AI alignment seeks to align large language models and other foundation models to the context-less and generic values of helpfulness, harmlessness, and honesty. Frontier model providers also strive to align their models with these values. In this paper, we motivate why we need to move beyond such a limited conception and propose three dimensions for doing so. The first scope of alignment is competence: knowledge, skills, or behaviors the model must possess to be useful for its intended purpose. The second scope of alignment is transience: either semantic or episodic depending on the context of use. The third scope of alignment is audience: either mass, public, small-group, or dyadic. At the end of the paper, we use the proposed framework to position some technologies and workflows that go beyond prevailing notions of alignment.

The exponential growth in the size and complexity of Large Language Models (LLMs) has introduced unprecedented challenges in their deployment and operational management. Traditional MLOps approaches often fail to efficiently handle the scale, resource requirements, and dynamic nature of these models. This research presents a novel framework that leverages Deep Neural Networks (DNNs) to optimize MLOps pipelines specifically for LLMs. Our approach introduces an intelligent system that automates deployment decisions, resource allocation, and pipeline optimization while maintaining optimal performance and cost efficiency. Through extensive experimentation across multiple cloud environments and deployment scenarios, we demonstrate significant improvements: 40% enhancement in resource utilization, 35% reduction in deployment latency, and 30% decrease in operational costs compared to traditional MLOps approaches. The framework's ability to adapt to varying workloads and automatically optimize deployment strategies represents a significant advancement in automated MLOps management for large-scale language models. Our framework introduces several novel components including a multi-stream neural architecture for processing heterogeneous operational metrics, an adaptive resource allocation system that continuously learns from deployment patterns, and a sophisticated deployment orchestration mechanism that automatically selects optimal strategies based on model characteristics and environmental conditions. The system demonstrates robust performance across various deployment scenarios, including multi-cloud environments, high-throughput production systems, and cost-sensitive deployments. Through rigorous evaluation using production workloads from multiple organizations, we validate our approach's effectiveness in reducing operational complexity while improving system reliability and cost efficiency.

Protecting Personally Identifiable Information (PII), such as names, is a critical requirement in learning technologies to safeguard student and teacher privacy and maintain trust. Accurate PII detection is an essential step toward anonymizing sensitive information while preserving the utility of educational data. Motivated by recent advancements in artificial intelligence, our study investigates the GPT-4o-mini model as a cost-effective and efficient solution for PII detection tasks. We explore both prompting and fine-tuning approaches and compare GPT-4o-mini's performance against established frameworks, including Microsoft Presidio and Azure AI Language. Our evaluation on two public datasets, CRAPII and TSCC, demonstrates that the fine-tuned GPT-4o-mini model achieves superior performance, with a recall of 0.9589 on CRAPII. Additionally, fine-tuned GPT-4o-mini significantly improves precision scores (a threefold increase) while reducing computational costs to nearly one-tenth of those associated with Azure AI Language. Furthermore, our bias analysis reveals that the fine-tuned GPT-4o-mini model consistently delivers accurate results across diverse cultural backgrounds and genders. The generalizability analysis using the TSCC dataset further highlights its robustness, achieving a recall of 0.9895 with minimal additional training data from TSCC. These results emphasize the potential of fine-tuned GPT-4o-mini as an accurate and cost-effective tool for PII detection in educational data. It offers robust privacy protection while preserving the data's utility for research and pedagogical analysis. Our code is available on GitHub: https://github.com/AnonJD/PrivacyAI

Currently, data and model size dominate the narrative in the training of super-large, powerful models. However, there has been a lack of exploration on the effect of other attributes of the training dataset on model performance. We hypothesize that dataset diversity can impact the performance of vision models. Our study shows positive correlations between test set accuracy and data diversity, providing an argument for furthering the research of dataset attributes beyond size. We analyzed pre-training and model-agnostic meta-learning methods on twelve popular visual datasets (e.g., Omniglot, CIFAR-FS, Aircraft) and five model configurations, including MAML variants with different numbers of inner gradient steps and supervised learning. We show moderate to strong positive correlations (R-squared: 0.15-0.42) between accuracy and data diversity and weaker but significant correlations (R-squared: ~0.2) between loss and diversity. These findings support our hypothesis and demonstrate a promising way for a deeper exploration of how formal data diversity influences model performance. This initial study highlights the potential of (Task2Vec) data diversity as a valuable measure in the rapidly evolving field of large-scale learning and emphasizes that understanding the dataset is key to building more powerful and generalizable models.

Many AI companies are training their large language models (LLMs) on data without the permission of the copyright owners. The permissibility of doing so varies by jurisdiction: in countries like the EU and Japan, this is allowed under certain restrictions, while in the United States, the legal landscape is more ambiguous. Regardless of the legal status, concerns from creative producers have led to several high-profile copyright lawsuits, and the threat of litigation is commonly cited as a reason for the recent trend towards minimizing the information shared about training datasets by both corporate and public interest actors. This trend in limiting data information causes harm by hindering transparency, accountability, and innovation in the broader ecosystem by denying researchers, auditors, and impacted individuals access to the information needed to understand AI models. While this could be mitigated by training language models on open access and public domain data, at the time of writing, there are no such models (trained at a meaningful scale) due to the substantial technical and sociological challenges in assembling the necessary corpus. These challenges include incomplete and unreliable metadata, the cost and complexity of digitizing physical records, and the diverse set of legal and technical skills required to ensure relevance and responsibility in a quickly changing landscape. Building towards a future where AI systems can be trained on openly licensed data that is responsibly curated and governed requires collaboration across legal, technical, and policy domains, along with investments in metadata standards, digitization, and fostering a culture of openness.

Automated interpretability pipelines generate natural language descriptions for the concepts represented by features in large language models (LLMs), such as plants or the first word in a sentence. These descriptions are derived using inputs that activate the feature, which may be a dimension or a direction in the model's representation space. However, identifying activating inputs is costly, and the mechanistic role of a feature in model behavior is determined both by how inputs cause a feature to activate and by how feature activation affects outputs. Using steering evaluations, we reveal that current pipelines provide descriptions that fail to capture the causal effect of the feature on outputs. To fix this, we propose efficient, output-centric methods for automatically generating feature descriptions. These methods use the tokens weighted higher after feature stimulation or the highest weight tokens after applying the vocabulary "unembedding" head directly to the feature. Our output-centric descriptions better capture the causal effect of a feature on model outputs than input-centric descriptions, but combining the two leads to the best performance on both input and output evaluations. Lastly, we show that output-centric descriptions can be used to find inputs that activate features previously thought to be "dead".

Large Language Models (LLMs) have significantly aided developers by generating or assisting in code writing, enhancing productivity across various tasks. While identifying incorrect code is often straightforward, detecting vulnerabilities in functionally correct code is more challenging, especially for developers with limited security knowledge, which poses considerable security risks of using LLM-generated code and underscores the need for robust evaluation benchmarks that assess both functional correctness and security. Current benchmarks like CyberSecEval and SecurityEval attempt to solve it but are hindered by unclear and impractical specifications, failing to assess both functionality and security accurately. To tackle these deficiencies, we introduce CWEval, a novel outcome-driven evaluation framework designed to enhance the evaluation of secure code generation by LLMs. This framework not only assesses code functionality but also its security simultaneously with high-quality task specifications and outcome-driven test oracles which provides high accuracy. Coupled with CWEval-bench, a multilingual, security-critical coding benchmark, CWEval provides a rigorous empirical security evaluation on LLM-generated code, overcoming previous benchmarks' shortcomings. Through our evaluations, CWEval reveals a notable portion of functional but insecure code produced by LLMs, and shows a serious inaccuracy of previous evaluations, ultimately contributing significantly to the field of secure code generation. We open-source our artifact at: https://github.com/Co1lin/CWEval .

Large language models excel at interpreting complex natural language instructions, enabling them to perform a wide range of tasks. In the life sciences, single-cell RNA sequencing (scRNA-seq) data serves as the "language of cellular biology", capturing intricate gene expression patterns at the single-cell level. However, interacting with this "language" through conventional tools is often inefficient and unintuitive, posing challenges for researchers. To address these limitations, we present InstructCell, a multi-modal AI copilot that leverages natural language as a medium for more direct and flexible single-cell analysis. We construct a comprehensive multi-modal instruction dataset that pairs text-based instructions with scRNA-seq profiles from diverse tissues and species. Building on this, we develop a multi-modal cell language architecture capable of simultaneously interpreting and processing both modalities. InstructCell empowers researchers to accomplish critical tasks--such as cell type annotation, conditional pseudo-cell generation, and drug sensitivity prediction--using straightforward natural language commands. Extensive evaluations demonstrate that InstructCell consistently meets or exceeds the performance of existing single-cell foundation models, while adapting to diverse experimental conditions. More importantly, InstructCell provides an accessible and intuitive tool for exploring complex single-cell data, lowering technical barriers and enabling deeper biological insights.