Previous research on LLM vulnerabilities often relied on nonsensical adversarial prompts, which were easily detectable by automated methods. We address this gap by focusing on human-readable adversarial prompts, a more realistic and potent threat. Our key contributions are situation-driven attacks leveraging movie scripts to create contextually relevant, human-readable prompts that successfully deceive LLMs, adversarial suffix conversion to transform nonsensical adversarial suffixes into meaningful text, and AdvPrompter with p-nucleus sampling, a method to generate diverse, human-readable adversarial suffixes, improving attack efficacy in models like GPT-3.5 and Gemma 7B. Our findings demonstrate that LLMs can be tricked by sophisticated adversaries into producing harmful responses with human-readable adversarial prompts and that there exists a scope for improvement when it comes to robust LLMs.

With the rise of foundation models, there is growing concern about their potential social impacts. Social science has a long history of studying the social impacts of transformative technologies in terms of pre-existing systems of power and how these systems are disrupted or reinforced by new technologies. In this position paper, we build on prior work studying the social impacts of earlier technologies to propose a conceptual framework studying foundation models as sociotechnical systems, incorporating social science expertise to better understand how these models affect systems of power, anticipate the impacts of deploying these models in various applications, and study the effectiveness of technical interventions intended to mitigate social harms. We advocate for an interdisciplinary and collaborative research paradigm between AI and social science across all stages of foundation model research and development to promote socially responsible research practices and use cases, and outline several strategies to facilitate such research.

Federated training methods have gained popularity for graph learning with applications including friendship graphs of social media sites and customer-merchant interaction graphs of huge online marketplaces. However, privacy regulations often require locally generated data to be stored on local clients. The graph is then naturally partitioned across clients, with no client permitted access to information stored on another. Cross-client edges arise naturally in such cases and present an interesting challenge to federated training methods, as training a graph model at one client requires feature information of nodes on the other end of cross-client edges. Attempting to retain such edges often incurs significant communication overhead, and dropping them altogether reduces model performance. In simpler models such as Graph Convolutional Networks, this can be fixed by communicating a limited amount of feature information across clients before training, but GATs (Graph Attention Networks) require additional information that cannot be pre-communicated, as it changes from training round to round. We introduce the Federated Graph Attention Network (FedGAT) algorithm for semi-supervised node classification, which approximates the behavior of GATs with provable bounds on the approximation error. FedGAT requires only one pre-training communication round, significantly reducing the communication overhead for federated GAT training. We then analyze the error in the approximation and examine the communication overhead and computational complexity of the algorithm. Experiments show that FedGAT achieves nearly the same accuracy as a GAT model in a centralised setting, and its performance is robust to the number of clients as well as data distribution.

The increasing demand for privacy-preserving data analytics in finance necessitates solutions for synthetic data generation that rigorously uphold privacy standards. We introduce DP-Fed-FinDiff framework, a novel integration of Differential Privacy, Federated Learning and Denoising Diffusion Probabilistic Models designed to generate high-fidelity synthetic tabular data. This framework ensures compliance with stringent privacy regulations while maintaining data utility. We demonstrate the effectiveness of DP-Fed-FinDiff on multiple real-world financial datasets, achieving significant improvements in privacy guarantees without compromising data quality. Our empirical evaluations reveal the optimal trade-offs between privacy budgets, client configurations, and federated optimization strategies. The results affirm the potential of DP-Fed-FinDiff to enable secure data sharing and robust analytics in highly regulated domains, paving the way for further advances in federated learning and privacy-preserving data synthesis.

There is a significant body of work looking at the ethical considerations of large language models (LLMs): critiquing tools to measure performance and harms; proposing toolkits to aid in ideation; discussing the risks to workers; considering legislation around privacy and security etc. As yet there is no work that integrates these resources into a single practical guide that focuses on LLMs; we attempt this ambitious goal. We introduce 'LLM Ethics Whitepaper', which we provide as an open and living resource for NLP practitioners, and those tasked with evaluating the ethical implications of others' work. Our goal is to translate ethics literature into concrete recommendations and provocations for thinking with clear first steps, aimed at computer scientists. 'LLM Ethics Whitepaper' distils a thorough literature review into clear Do's and Don'ts, which we present also in this paper. We likewise identify useful toolkits to support ethical work. We refer the interested reader to the full LLM Ethics Whitepaper, which provides a succinct discussion of ethical considerations at each stage in a project lifecycle, as well as citations for the hundreds of papers from which we drew our recommendations. The present paper can be thought of as a pocket guide to conducting ethical research with LLMs.

The extensive application of Large Language Models (LLMs) in generative coding tasks has raised concerns due to their high computational demands and energy consumption. Unlike previous structural pruning methods designed for classification models that deal with lowdimensional classification logits, generative Code LLMs produce high-dimensional token logit sequences, making traditional pruning objectives inherently limited. Moreover, existing single component pruning approaches further constrain the effectiveness when applied to generative Code LLMs. In response, we propose Flab-Pruner, an innovative unified structural pruning method that combines vocabulary, layer, and Feed-Forward Network (FFN) pruning. This approach effectively reduces model parameters while maintaining performance. Additionally, we introduce a customized code instruction data strategy for coding tasks to enhance the performance recovery efficiency of the pruned model. Through extensive evaluations on three state-of-the-art Code LLMs across multiple generative coding tasks, the results demonstrate that Flab-Pruner retains 97% of the original performance after pruning 22% of the parameters and achieves the same or even better performance after post-training. The pruned models exhibit significant improvements in storage, GPU usage, computational efficiency, and environmental impact, while maintaining well robustness. Our research provides a sustainable solution for green software engineering and promotes the efficient deployment of LLMs in real-world generative coding intelligence applications.

This paper presents a new approach to fine-tuning OpenAI's Whisper model for low-resource languages by introducing a novel data generation method that converts sentence-level data into a long-form corpus, using Swiss German as a case study. Non-sentence-level data, which could improve the performance of long-form audio, is difficult to obtain and often restricted by copyright laws. Our method bridges this gap by transforming more accessible sentence-level data into a format that preserves the model's ability to handle long-form audio and perform segmentation without requiring non-sentence-level data. Our data generation process improves performance in several real-world applications and leads to the development of a new state-of-the-art speech-to-text (STT) model for Swiss German. We compare our model with a non-fine-tuned Whisper and our previous state-of-the-art Swiss German STT models, where our new model achieves higher BLEU scores. Our results also indicate that the proposed method is adaptable to other low-resource languages, supported by written guidance and code that allows the creation of fine-tuned Whisper models, which keep segmentation capabilities and allow the transcription of longer audio files using only sentence-level data with high quality.

The current study investigates possible neural mechanisms underling autonomous shifts between focus state and mind-wandering by conducting model simulation experiments. On this purpose, we modeled perception processes of continuous sensory sequences using our previous proposed variational RNN model which was developed based on the free energy principle. The current study extended this model by introducing an adaptation mechanism of a meta-level parameter, referred to as the meta-prior $\mathbf{w}$, which regulates the complexity term in the free energy. Our simulation experiments demonstrated that autonomous shifts between focused perception and mind-wandering take place when $\mathbf{w}$ switches between low and high values associated with decrease and increase of the average reconstruction error over the past window. In particular, high $\mathbf{w}$ prioritized top-down predictions while low $\mathbf{w}$ emphasized bottom-up sensations. This paper explores how our experiment results align with existing studies and highlights their potential for future research.

But it's now clear that the case--filed more than two years before the generative AI boom began--was the first strike in a much larger war between content publishers and artificial intelligence companies now unfolding in courts across the country. The outcome could make, break, or reshape the information ecosystem and the entire AI industry--and in doing so, impact just about everyone across the internet. The plaintiffs include individual authors like Sarah Silverman and Ta Nehisi-Coates, visual artists, media companies like The New York Times, and music-industry giants like Universal Music Group. This wide variety of rights holders are alleging that AI companies have used their work to train what are often highly lucrative and powerful AI models in a manner that is tantamount to theft. Nearly every major generative AI company has been pulled into this legal fight, including OpenAI, Meta, Microsoft, Google, Anthropic, and Nvidia.

Fox News chief political anchor Bret Baier has the latest on regulatory uncertainty amid AI development on'Special Report.' Congress's bipartisan task force on artificial intelligence (AI) released its long-anticipated report this week, detailing strategies for how the U.S. can protect itself against emerging AI-related threats while ensuring the nation remains a leader in innovation within this rapidly evolving sector. Responses to the report, which sought to strike a "flexible sectoral regulatory framework," were positive and with mixed concerns. "The Task Force report offers a thoughtful and forward-thinking framework that balances AI's transformative economic potential with the imperative to address legitimate safety concerns," said Dr. Vahid Behzadan, a professor in the computer science department at the University of New Haven. "That said, there's still work to be done."