We conducted controlled experimental bias audits for four versions of ChatGPT, which we asked to recommend an opening offer in salary negotiations for a new hire. We submitted 98,800 prompts to each version, systematically varying the employee's gender, university, and major, and tested prompts in voice of each side of the negotiation: the employee versus employer. We find ChatGPT as a multi-model platform is not robust and consistent enough to be trusted for such a task. We observed statistically significant salary offers when varying gender for all four models, although with smaller gaps than for other attributes tested. The largest gaps were different model versions and between the employee- vs employer-voiced prompts. We also observed substantial gaps when varying university and major, but many of the biases were not consistent across model versions. We tested for fictional and fraudulent universities and found wildly inconsistent results across cases and model versions. We make broader contributions to the AI/ML fairness literature. Our scenario and our experimental design differ from mainstream AI/ML auditing efforts in key ways. Bias audits typically test discrimination for protected classes like gender, which we contrast with testing non-protected classes of university and major. Asking for negotiation advice includes how aggressive one ought to be in a negotiation relative to known empirical salary distributions and scales, which is a deeply contextual and personalized task that has no objective ground truth to validate. These results raise concerns for the specific model versions we tested and ChatGPT as a multi-model platform in continuous development. Our epistemology does not permit us to definitively certify these models as either generally biased or unbiased on the attributes we test, but our study raises matters of concern for stakeholders to further investigate.

Compressive adaptation approaches, such as QLoRA, are widely popular alternatives for reducing memory requirements during fine-tuning of large language models (LLMs) while producing models capable of handling various downstream tasks. The key idea is to employ a "two-tower" architecture: compressing pretrained LLM parameters into compact representations and fine-tuning the additive full-precision adapter, which typically has few tunable parameters in low-rank format. However, the strict algebraic assumptions, such as low-rank assumption, and the complexity of composing two-tower architectures are some of the known shortcomings, resulting in a poor accuracy-efficiency trade-off. In response to these known limitations, we propose SpaLLM (Sketched Parameter Adaptation of LLMs), a novel compressive adaptation approach for LLMs. This method is also the first to illustrate parameter-sharing compression methods for LLM finetuning, which, unlike QLoRA, are free from strict low-rank algebraic assumptions on adapters. This approach simplifies LLMs' compressive adaptation workflow, potentially improves multi-user serving efficiency, and delivers significantly better accuracy for both natural language understanding and generation tasks. Moreover, by avoiding the "two-tower" architecture, our framework only requires one compressed matrix multiplication per layer during inference, demonstrating superior inference efficiency compared to previous methods. Recent advancements in Large Language Models (LLMs) have demonstrated exceptional performance in Natural Language Processing (NLP), enabling a broad spectrum of downstream applications. LLMs have demonstrated impressive generalization abilities across many downstream tasks in a zero-shot manner. However, compared to training-free methods such as in-context learning (Dong et al., 2022; Rubin et al., 2021) and few-shot prompting (Brown, 2020; Song et al., 2023), fine-tuning on these LLMs is often the ideal method to achieve optimal performance on a specific downstream task (Ding et al., 2023). Clearly, full-precision fine-tuning on these LLMs are often impractical due to the massive requirement of high-performance computing devices such as GPUs. As a result, Parameter-Efficient Fine-Tuning methods (PEFT), such as Low-Rank Adaptation (LoRA) (Hu et al., 2022), emerged as a less resource-intensive approach to fine-tuning while achieving reasonable Clearly, there is a trade-off between accuracy and efficiency.

Document understanding is a challenging task to process and comprehend large amounts of textual and visual information. Recent advances in Large Language Models (LLMs) have significantly improved the performance of this task. However, existing methods typically focus on either plain text or a limited number of document images, struggling to handle long PDF documents with interleaved text and images, especially in academic papers. In this paper, we introduce PDF-WuKong, a multimodal large language model (MLLM) which is designed to enhance multimodal question-answering (QA) for long PDF documents. PDF-WuKong incorporates a sparse sampler that operates on both text and image representations, significantly improving the efficiency and capability of the MLLM. The sparse sampler is integrated with the MLLM's image encoder and selects the paragraphs or diagrams most pertinent to user queries for processing by the language model. To effectively train and evaluate our model, we construct PaperPDF, a dataset consisting of a broad collection of academic papers sourced from arXiv, multiple strategies are proposed to generate automatically 1M QA pairs along with their corresponding evidence sources. Experimental results demonstrate the superiority and high efficiency of our approach over other models on the task of long multimodal PDF understanding, surpassing proprietary products by an average of 8.6% on F1. Our code and dataset will be released at https://github.com/yh-hust/PDF-Wukong.

Foundation Models (FMs) display exceptional performance in tasks such as natural language processing and are being applied across a growing range of disciplines. Although typically trained on large public datasets, FMs are often fine-tuned or integrated into Retrieval-Augmented Generation (RAG) systems, which rely on private data. This access, along with their size and costly training, heightens the risk of intellectual property theft. Moreover, multimodal FMs may expose sensitive information. In this work, we examine the FM threat model and discuss the practicality and comprehensiveness of various approaches for securing against them, such as ML-based methods and trusted execution environments (TEEs). We demonstrate that TEEs offer an effective balance between strong security properties, usability, and performance. Specifically, we present a solution achieving less than 10\% overhead versus bare metal for the full Llama2 7B and 13B inference pipelines running inside \intel\ SGX and \intel\ TDX. We also share our configuration files and insights from our implementation. To our knowledge, our work is the first to show the practicality of TEEs for securing FMs.

A vast amount of textual data is added to the internet daily, making utilization and interpretation of such data difficult and cumbersome. As a result, automatic text summarization is crucial for extracting relevant information, saving precious reading time. Although many transformer-based models excel in summarization, they are constrained by their input size, preventing them from processing texts longer than their context size. This study introduces three novel algorithms that allow any LLM to efficiently overcome its input size limitation, effectively utilizing its full potential without any architectural modifications. We test our algorithms on texts with more than 70,000 words, and our experiments show a significant increase in BERTScore with competitive ROUGE scores.

Universities are increasingly adopting data-driven strategies to enhance student success, with AI applications like Learning Analytics (LA) and Predictive Learning Analytics (PLA) playing a key role in identifying at-risk students, personalising learning, supporting teachers, and guiding educational decision-making. However, concerns are rising about potential harms these systems may pose, such as algorithmic biases leading to unequal support for minority students. While many have explored the need for Responsible AI in LA, existing works often lack practical guidance for how institutions can operationalise these principles. In this paper, we propose a novel Responsible AI framework tailored specifically to LA in Higher Education (HE). We started by mapping 11 established Responsible AI frameworks, including those by leading tech companies, to the context of LA in HE. This led to the identification of seven key principles such as transparency, fairness, and accountability. We then conducted a systematic review of the literature to understand how these principles have been applied in practice. Drawing from these findings, we present a novel framework that offers practical guidance to HE institutions and is designed to evolve with community input, ensuring its relevance as LA systems continue to develop.

In sensitive domains, such as legal or medial domains, the correctness of information given to users is critical. To address this, the recently introduced task Conversational Tree Search (CTS) provides a graph-based framework for controllable task-oriented dialog in sensitive domains. However, a big drawback of state-of-the-art CTS agents is their long training time, which is especially problematic as a new agent must be trained every time the associated domain graph is updated. The goal of this paper is to eliminate the need for training CTS agents altogether. To achieve this, we implement a novel LLM-based method for zero-shot, controllable CTS agents. We show that these agents significantly outperform state-of-the-art CTS agents (p<0.0001; Barnard Exact test) in simulation. This generalizes to all available CTS domains. Finally, we perform user evaluation to test the agent performance in the wild, showing that our policy significantly (p<0.05; Barnard Exact) improves task-success compared to the state-of-the-art Reinforcement Learning-based CTS agent.

The success of modern Machine Learning (ML) systems depends on the quality and quantity of data used for training, which directly influences model performance and generalization capabilities. To this aim, high-quality, diverse, and representative datasets are essential for accurate and unbiased predictions. For instance, insufficient or biased data can lead to poor model performance, inaccuracies, and unintended consequences. Ethical and legal aspects are critical, too.

Unlike regular models, these very simple models--often referred to as toy models--are not required to be linked to the real world through structural similarity or resemblance relations. They are not meant to be approximations of the target world system, and in some cases, they are not even required to be representational. In semantic terms, they do not accurately map onto their targets. Despite these limitations, they are still useful in understanding theoretical concepts and possible configurations of the target system. Paradigmatic examples of toy models include Boyle's law and the Ising model in physics, the Lotka-Volterra model in population ecology, and the Schelling model in the social sciences (Weisberg, 2013). In recent years, philosophers of science have become interested in toy models (Grüne-Yanoff, 2009; Luczak, 2017; Reutlinger et al., 2018; Frigg & Nguyen, 2017; Nguyen, 2020). The main purpose of this literature is to explore the nature of these models and examine how they perform their epistemic function. Despite lacking the regular descriptive and predictive features of full-scale scientific models, they often offer an elementary understanding of a phenomenon. Their definitions of "toy model" differ as well as their assessment of the importance of representation in modelling generally, but they all agree that toy models play an important epistemic role in scientific research, exploration, and pedagogy.

To mitigate societal biases implicitly encoded in recent successful pretrained language models, a diverse array of approaches have been proposed to encourage model fairness, focusing on prompting, data augmentation, regularized finetuning, and more. Despite the development, it is nontrivial to reach a principled understanding of fairness and an effective algorithm that can consistently debias language models. In this work, by rigorous evaluations of Neural Collapse - a learning phenomenon happen in last-layer representations and classifiers in deep networks - on fairness-related words, we find that debiased language models exhibit collapsed alignment between token representations and word embeddings. More importantly, this observation inspires us to design a principled fine-tuning method that can effectively improve fairness in a wide range of debiasing methods, while still preserving the performance of language models on standard natural language understanding tasks. The rise of pre-trained language models (PLMs) has revolutionized natural language processing, greatly enhancing tasks like reasoning and prediction by harnessing the semantic richness of language data. Despite their effectiveness, these models, trained on extensive corpora, often reflect and even intensify societal biases in their training datasets. Such biases manifest in the association of demographic groups with specific roles or capabilities, affecting fairness in applications ranging from legal analytics to hiring processes [49; 12; 38; 2; 52; 3; 7]. Thus, it is crucial to address and mitigate these biases to prevent discriminatory practices in downstream applications [70; 64; 46].