LLM watermarking has attracted attention as a promising way to detect AI-generated content, with some works suggesting that current schemes may already be fit for deployment. In this work we dispute this claim, identifying watermark stealing (WS) as a fundamental vulnerability of these schemes. We show that querying the API of the watermarked LLM to approximately reverse-engineer a watermark enables practical spoofing attacks, as hypothesized in prior work, but also greatly boosts scrubbing attacks, which was previously unnoticed. We are the first to propose an automated WS algorithm and use it in the first comprehensive study of spoofing and scrubbing in realistic settings. We show that for under $50 an attacker can both spoof and scrub state-of-the-art schemes previously considered safe, with average success rate of over 80%. Our findings challenge common beliefs about LLM watermarking, stressing the need for more robust schemes. We make all our code and additional examples available at https://watermark-stealing.org.

Despite the remarkable generative capabilities of language models in producing naturalistic language, their effectiveness on explicit manipulation and generation of linguistic structures remain understudied. In this paper, we investigate the task of generating new sentences preserving a given semantic structure, following the FrameNet formalism. We propose a framework to produce novel frame-semantically annotated sentences following an overgenerate-and-filter approach. Our results show that conditioning on rich, explicit semantic information tends to produce generations with high human acceptance, under both prompting and finetuning. Our generated frame-semantic structured annotations are effective at training data augmentation for frame-semantic role labeling in low-resource settings; however, we do not see benefits under higher resource settings. Our study concludes that while generating high-quality, semantically rich data might be within reach, the downstream utility of such generations remains to be seen, highlighting the outstanding challenges with automating linguistic annotation tasks.

As Large Language Models (LLMs) continue to exhibit remarkable performance in natural language understanding tasks, there is a crucial need to measure their ability for human-like multi-step logical reasoning. Existing logical reasoning evaluation benchmarks often focus primarily on simplistic single-step or multi-step reasoning with a limited set of inference rules. Furthermore, the lack of datasets for evaluating non-monotonic reasoning represents a crucial gap since it aligns more closely with human-like reasoning. To address these limitations, we propose Multi-LogiEval, a comprehensive evaluation dataset encompassing multi-step logical reasoning with various inference rules and depths. Multi-LogiEval covers three logic types--propositional, first-order, and non-monotonic--consisting of more than 30 inference rules and more than 60 of their combinations with various depths. Leveraging this dataset, we conduct evaluations on a range of LLMs including GPT-4, ChatGPT, Gemini-Pro, Yi, Orca, and Mistral, employing a zero-shot chain-of-thought. Experimental results show that there is a significant drop in the performance of LLMs as the reasoning steps/depth increases (average accuracy of ~68% at depth-1 to ~43% at depth-5). We further conduct a thorough investigation of reasoning chains generated by LLMs which reveals several important findings. We believe that Multi-LogiEval facilitates future research for evaluating and enhancing the logical reasoning ability of LLMs. Data is available at https://github.com/Mihir3009/Multi-LogiEval.

While large language models (LLMs) excel in many domains, their complexity and scale challenge deployment in resource-limited environments. Current compression techniques, such as parameter pruning, often fail to effectively utilize the knowledge from pruned parameters. To address these challenges, we propose Manifold-Based Knowledge Alignment and Layer Merging Compression (MKA), a novel approach that uses manifold learning and the Normalized Pairwise Information Bottleneck (NPIB) measure to merge similar layers, reducing model size while preserving essential performance. We evaluate MKA on multiple benchmark datasets and various LLMs. Our findings show that MKA not only preserves model performance but also achieves substantial compression ratios, outperforming traditional pruning methods. Moreover, when coupled with quantization, MKA delivers even greater compression. Specifically, on the MMLU dataset using the Llama3-8B model, MKA achieves a compression ratio of 43.75% with a minimal performance decrease of only 2.82\%. The proposed MKA method offers a resource-efficient and performance-preserving model compression technique for LLMs.

We investigate the abilities of a representative set of Large language Models (LLMs) to reason about cardinal directions (CDs). To do so, we create two datasets: the first, co-created with ChatGPT, focuses largely on recall of world knowledge about CDs; the second is generated from a set of templates, comprehensively testing an LLM's ability to determine the correct CD given a particular scenario. The templates allow for a number of degrees of variation such as means of locomotion of the agent involved, and whether set in the first , second or third person. Even with a temperature setting of zero, Our experiments show that although LLMs are able to perform well in the simpler dataset, in the second more complex dataset no LLM is able to reliably determine the correct CD, even with a temperature setting of zero.

Since the advent of GPT, large language models (LLMs) have brought about revolutionary advancements in all walks of life. As a superior natural language processing (NLP) technology, LLMs have consistently achieved state-of-the-art performance on numerous areas. However, LLMs are considered to be general-purpose models for NLP tasks, which may encounter challenges when applied to complex tasks in specialized fields such as optical networks. In this study, we propose a framework of LLM-empowered optical networks, facilitating intelligent control of the physical layer and efficient interaction with the application layer through an LLM-driven agent (AI-Agent) deployed in the control layer. The AI-Agent can leverage external tools and extract domain knowledge from a comprehensive resource library specifically established for optical networks. This is achieved through user input and well-crafted prompts, enabling the generation of control instructions and result representations for autonomous operation and maintenance in optical networks. To improve LLM's capability in professional fields and stimulate its potential on complex tasks, the details of performing prompt engineering, establishing domain knowledge library, and implementing complex tasks are illustrated in this study. Moreover, the proposed framework is verified on two typical tasks: network alarm analysis and network performance optimization. The good response accuracies and sematic similarities of 2,400 test situations exhibit the great potential of LLM in optical networks.

This research introduces a novel approach for achieving semantic interoperability in digital twins and assisting the creation of Asset Administration Shell (AAS) as digital twin model within the context of Industry 4.0. The foundational idea of our research is that the communication based on semantics and the generation of meaningful textual data are directly linked, and we posit that these processes are equivalent if the exchanged information can be serialized in text form. Based on this, we construct a "semantic node" data structure in our research to capture the semantic essence of textual data. Then, a system powered by large language models is designed and implemented to process the "semantic node" and generate standardized digital twin models from raw textual data collected from datasheets describing technical assets. Our evaluation demonstrates an effective generation rate of 62-79%, indicating a substantial proportion of the information from the source text can be translated error-free to the target digital twin instance model with the generative capability of large language models. This result has a direct application in the context of Industry 4.0, and the designed system is implemented as a data model generation tool for reducing the manual effort in creating AAS model. In our evaluation, a comparative analysis of different LLMs and an in-depth ablation study of Retrieval-Augmented Generation (RAG) mechanisms provide insights into the effectiveness of LLM systems for interpreting technical concepts and translating data. Our findings emphasize LLMs' capability to automate AAS instance creation and contribute to the broader field of semantic interoperability for digital twins in industrial applications. The prototype implementation and evaluation results are presented on our GitHub Repository: https://github.com/YuchenXia/AASbyLLM.

We are increasingly subjected to the power of AI authorities. As AI decisions become inescapable, entering domains such as healthcare, education, and law, we must confront a vital question: how can we ensure AI systems have the legitimacy necessary for effective governance? This essay argues that to secure AI legitimacy, we need methods that engage the public in designing and constraining AI systems, ensuring these technologies reflect the community's shared values. Constitutional AI, proposed by Anthropic, represents a step towards this goal, offering a model for democratic control of AI. However, while Constitutional AI's commitment to hardcoding explicit principles into AI models enhances transparency and accountability, it falls short in two crucial aspects: addressing the opacity of individual AI decisions and fostering genuine democratic legitimacy. To overcome these limitations, this essay proposes "Public Constitutional AI." This approach envisions a participatory process where diverse stakeholders, including ordinary citizens, deliberate on the principles guiding AI development. The resulting "AI Constitution" would carry the legitimacy of popular authorship, grounding AI governance in the public will. Furthermore, the essay proposes "AI Courts" to develop "AI case law," providing concrete examples for operationalizing constitutional principles in AI training. This evolving combination of constitutional principles and case law aims to make AI governance more responsive to public values. By grounding AI governance in deliberative democratic processes, Public Constitutional AI offers a path to imbue automated authorities with genuine democratic legitimacy, addressing the unique challenges posed by increasingly powerful AI systems while ensuring their alignment with the public interest.

Recently, there has been increasing interest in applying large language models (LLMs) as zero-shot passage rankers. However, few studies have explored how to select appropriate in-context demonstrations for the passage ranking task, which is the focus of this paper. Previous studies mainly apply a demonstration retriever to retrieve demonstrations and use top-$k$ demonstrations for in-context learning (ICL). Although effective, this approach overlooks the dependencies between demonstrations, leading to inferior performance of few-shot ICL in the passage ranking task. In this paper, we formulate the demonstration selection as a \textit{retrieve-then-rerank} process and introduce the DemoRank framework. In this framework, we first use LLM feedback to train a demonstration retriever and construct a novel dependency-aware training samples to train a demonstration reranker to improve few-shot ICL. The construction of such training samples not only considers demonstration dependencies but also performs in an efficient way. Extensive experiments demonstrate DemoRank's effectiveness in in-domain scenarios and strong generalization to out-of-domain scenarios. Our codes are available at~\url{https://github.com/8421BCD/DemoRank}.

Prompting Large Language Models (LLMs) has created new and interesting means for classifying textual data. While evaluating and remediating group fairness is a well-studied problem in classifier fairness literature, some classical approaches (e.g., regularization) do not carry over, and some new opportunities arise (e.g., prompt-based remediation). We measure fairness of LLM-based classifiers on a toxicity classification task, and empirically show that prompt-based classifiers may lead to unfair decisions. We introduce several remediation techniques and benchmark their fairness and performance trade-offs. We hope our work encourages more research on group fairness in LLM-based classifiers.