We examine the assumption that hidden-state vectors of recurrent neural networks (RNNs) tend to form clusters of semantically similar vectors, which we dub the clustering hypothesis. While this hypothesis has been assumed in RNN analyses in recent years, its validity has not been studied thoroughly on modern RNN architectures. We first consider RNNs that were trained to recognize regular languages. This enables us to draw on perfect ground-truth automata in our evaluation, against which we can compare the RNN’s accuracy and the distribution of the hidden-state vectors. Then, we consider context-free languages to examine if RNN states form clusters for more expressive languages.For our analysis, we fit (generalized) linear models to classify RNN states into automata states and we apply different unsupervised clustering techniques. With a new ambiguity score, derived from information entropy, we measure how well an abstraction function maps the hidden state vectors to abstract clusters. Our evaluation supports the validity of the clustering hypothesis for regular languages, especially if RNNs are well-trained, i.e., clustering techniques succeed in finding clusters of similar state vectors. However, the clustering accuracy decreases substantially for context-free languages. This suggests that clustering is not a reliable abstraction technique for RNNs used in tasks like natural language processing.
论文检索
输入标题、作者或关键词，从 1,915 篇学术成果中精准定位
The focus of language model evaluation has transitioned towards reasoning and knowledge-intensive tasks, driven by advancements in pretraining large models. While state-of-the-art models are partially trained on large Arabic texts, evaluating their performance in Arabic remains challenging due to the limited availability of relevant datasets. To bridge this gap, we present ArabicMMLU, the first multi-task language understanding benchmark for the Arabic language, sourced from school exams across diverse educational levels in different countries spanning North Africa, the Levant, and the Gulf regions. Our data comprises 40 tasks and 14,575 multiple-choice questions in Modern Standard Arabic (MSA) and is carefully constructed by collaborating with native speakers in the region. Our comprehensive evaluations of 35 models reveal substantial room for improvement, particularly among the best open-source models. Notably, BLOOMZ, mT0, LLama2, and Falcon struggle to achieve a score of 50%, while even the top-performing Arabic-centric model only achieves a score of 62.3%.
In this paper, we investigate the phenomena of “selection biases” in Large Language Models (LLMs), focusing on problems where models are tasked with choosing the optimal option from an ordered sequence. We delve into biases related to option order and token usage, which significantly impact LLMs’ decision-making processes. We also quantify the impact of these biases through an extensive empirical analysis across multiple models and tasks. Furthermore, we propose mitigation strategies to enhance model performance. Our key contributions are threefold: 1) Precisely quantifying the influence of option order and token on LLMs, 2) Developing strategies to mitigate the impact of token and order sensitivity to enhance robustness, and 3) Offering a detailed analysis of sensitivity across models and tasks, which informs the creation of more stable and reliable LLM applications for selection problems.
Language Models (LMs) acquire factual knowledge during pre-training and store it in the parameters, which can be valuable for downstream tasks. As world evolves, some facts may be incorrectly induced or become obsolete over time. Various model editing methods have been proposed to modify specific examples in LMs. However, existing training-based methods still suffer from sub-optimal locality, where irrelevant neighborhood examples can be adversely influenced. Model’s gradients are still struggling to identify the appropriate direction when updating the parameters. To address this issue, we find that directing the hidden state of the edit example towards spaces where semantics are sparse tends to help preserve the semantics of irrelevant neighborhood examples. Based on this hypothesis, we propose a novel metric, named SSS, to evaluate the degree of sparsity around a sentence embedding in the semantic space without any human or machine annotation. Subsequently, we incorporate SSS into the original loss function of the existing training-based methods to enhance locality. Experiments conducted on two datasets across various models demonstrate that SSS is effective in improving both locality and reasoning capability.
Social media advertising offers a platform for fossil fuel value chain companies and their agents to reinforce their narratives, often emphasizing economic, labor market, and energy security benefits to promote oil and gas policy and products. Whether such narratives can be detected automatically and the extent to which the cost of human annotation can be reduced is our research question. We introduce a task of classifying narratives into seven categories, based on existing definitions and data.Experiments showed that RoBERTa-large outperforms other methods, while GPT-4 Turbo can serve as a viable annotator for the task, thereby reducing human annotation costs. Our findings and insights provide guidance to automate climate-related ad analysis and lead to more scalable ad scrutiny.
Improving the Robustness of Distantly-Supervised Named Entity Recognition via Uncertainty-Aware Teacher Learning and Student-Student Collaborative Learning
PDF ↗Distantly-Supervised Named Entity Recognition (DS-NER) effectively alleviates the burden of annotation, but meanwhile suffers from the label noise. Recent works attempt to adopt the teacher-student framework to gradually refine the training labels and improve the overall robustness. However, we argue that these teacher-student methods achieve limited performance because the poor calibration of the teacher network produces incorrectly pseudo-labeled samples, leading to error propagation. Therefore, we attempt to mitigate this issue by proposing: (1) Uncertainty-Aware Teacher Learning that leverages the prediction uncertainty to reduce the number of incorrect pseudo labels in the self-training stage; (2) Student-Student Collaborative Learning that allows the transfer of reliable labels between two student networks instead of indiscriminately relying on all pseudo labels from its teacher. This approach further enables a full exploration of mislabeled samples rather than simply filtering unreliable pseudo-labeled samples. We evaluate our proposed method on five DS-NER datasets, demonstrating that our method is superior to the state-of-the-art DS-NER denoising methods.
Thinking about how to extract: Energizing LLMs’ emergence capabilities for document-level event argument extraction
PDF ↗There are two key challenges remaining for the document-level event argument extraction (D-EAE) tasks: key feature forgetting and cross-event argument confusion. The emergence capability of large language models (LLMs) holds promise for solving the above two challenges. In this paper, we propose a document-level event argument extraction method based on guided summarization and reasoning (EAESR), which leverages the emergence capabilities of LLMs to highlight key event information and to clarify the explicit and implicit association between multiple events. Specifically, we generate document summarization information that shorten the length of the event context while preserving the key event features. In addition, we generate inter-event reasoning information, which helps EAESR make sense of the correlations between events and reduces their dependence on the event context, especially to better cope with the few-shot D-EAE task. Then, we obtain named entity information to enable EAESR to learn argument boundary features to improve the sensitivity of its argument boundary recognition. Eventually, we fused the above features and sentence features to make EAESR have summarizing and reasoning capabilities simultaneously. Extensive experiments on WIKIEVENTS and RAMS have shown that EAESR achieves a new state-of-the-art that outperforms the baseline models by 1.3% F1 and 1.6% F1, respectively, and averages 11% F1 in few-shot settings.
Subtle Signatures, Strong Shields: Advancing Robust and Imperceptible Watermarking in Large Language Models
PDF ↗The widespread adoption of Large Language Models (LLMs) has led to an increase in AI-generated text on the Internet, presenting a crucial challenge to differentiate AI-created content from human-written text. This challenge is critical to prevent issues of authenticity, trust, and potential copyright violations. Current research focuses on watermarking LLM-generated text, but traditional techniques struggle to balance robustness with text quality. We introduce a novel watermarking approach, Robust and Imperceptible Watermarking (RIW) for LLMs, which leverages token prior probabilities to improve detectability and maintain watermark imperceptibility. RIW methodically embeds watermarks by partitioning selected tokens into two distinct groups based on their prior probabilities and employing tailored strategies for each group. In the detection stage, the RIW method employs the ‘voted z-test’ to provide a statistically robust framework to identify the presence of a watermark accurately. The effectiveness of RIW is evaluated across three key dimensions: success rate, text quality, and robustness against removal attacks. Our experimental results on various LLMs, including GPT2-XL, OPT-1.3B, and LLaMA2-7B, indicate that RIW surpasses existing models, and also exhibits increased robustness against various attacks and good imperceptibility, thus promoting the responsible use of LLMs.
The potential for improvements brought by Large Language Models (LLMs) in Text-to-SQL systems is mostly assessed on monolingual English datasets. However, LLMs’ performance for other languages remains vastly unexplored. In this work, we release the StatBot.Swiss dataset, the first bilingual benchmark for evaluating Text-to-SQL systems based on real-world applications. The StatBot.Swiss dataset contains 455 natural language/SQL-pairs over 35 big databases with varying level of complexity for both English and German.We evaluate the performance of state-of-the-art LLMs such as GPT-3.5-Turbo and mixtral-8x7b-instruct for the Text-to-SQL translation task using an in-context learning approach. Our experimental analysis illustrates that current LLMs struggle to generalize well in generating SQL queries on our novel bilingual dataset.
Before Generation, Align it! A Novel and Effective Strategy for Mitigating Hallucinations in Text-to-SQL Generation
PDF ↗Large Language Models (LLMs) driven by In-Context Learning (ICL) have significantly improved the performance of text-to-SQL. Previous methods generally employ a two-stage reasoning framework, namely 1) schema linking and 2) logical synthesis, making the framework not only effective but also interpretable. Despite these advancements, the inherent bad nature of the generalization of LLMs often results in hallucinations, which limits the full potential of LLMs. In this work, we first identify and categorize the common types of hallucinations at each stage in text-to-SQL. We then introduce a novel strategy, Task Alignment (TA), designed to mitigate hallucinations at each stage. TA encourages LLMs to take advantage of experiences from similar tasks rather than starting the tasks from scratch. This can help LLMs reduce the burden of generalization, thereby mitigating hallucinations effectively. We further propose TA-SQL, a text-to-SQL framework based on this strategy. The experimental results and comprehensive analysis demonstrate the effectiveness and robustness of our framework. Specifically, it enhances the performance of the GPT-4 baseline by 21.23% relatively on BIRD dev and it yields significant improvements across six models and four mainstream, complex text-to-SQL benchmarks.
Can We Continually Edit Language Models? On the Knowledge Attenuation in Sequential Model Editing
PDF ↗Model editing has become a promising method for precisely and effectively updating knowledge in language models. In this paper, we investigate knowledge attenuation, in which the retention of updated knowledge within the language model decreases as the number of edits increases after sequential editing. Through empirical study, we discovered that existing editing methods generally suffer from knowledge attenuation. We attribute this phenomenon to two aspects: (1) redundant parameters interference and (2) update weight disentanglement. To this end, we propose the AdaPLE method. It not only mitigates the knowledge attenuation issue but also improves the performance on existing benchmarks. To the best of our knowledge, we are the first to investigate the cause and mitigation of knowledge attenuation in sequential LLM editing.
Although model editing has shown promise in revising knowledge in Large Language Models (LLMs), its impact on the inherent capabilities of LLMs is often overlooked. In this work, we reveal a critical phenomenon: even a single edit can trigger model collapse, manifesting as significant performance degradation in various benchmark tasks. However, benchmarking LLMs after each edit, while necessary to prevent such collapses, is impractically time-consuming and resource-intensive. To mitigate this, we propose using perplexity as a surrogate metric, validated by extensive experiments demonstrating changes in an edited model’s perplexity are strongly correlated with its downstream task performances. We further conduct an in-depth study on sequential editing, a practical setting for real-world scenarios, across various editing methods and LLMs, focusing on hard cases from our previous single edit studies. The results indicate that nearly all examined editing methods result in model collapse after only few edits. To facilitate further research, we have utilized GPT-3.5 to develop a new dataset, HardEdit, based on those hard cases. This dataset aims to establish the foundation for pioneering research in reliable model editing and the mechanisms underlying editing-induced model collapse. We hope this work can draw the community’s attention to the potential risks inherent in model editing practices.
Educational materials such as survey articles in specialized fields like computer science traditionally require tremendous expert inputs and are therefore expensive to create and update. Recently, Large Language Models (LLMs) have achieved significant success across various general tasks. However, their effectiveness and limitations in the education domain are yet to be fully explored. In this work, we examine the proficiency of LLMs in generating succinct survey articles specific to the niche field of NLP in computer science, focusing on a curated list of 99 topics. Automated benchmarks reveal that GPT-4 surpasses its predecessors, inluding GPT-3.5, PaLM2, and LLaMa2 by margins ranging from 2% to 20% in comparison to the established ground truth. We compare both human and GPT-based evaluation scores and provide in-depth analysis. While our findings suggest that GPT-created surveys are more contemporary and accessible than human-authored ones, certain limitations were observed. Notably, GPT-4, despite often delivering outstanding content, occasionally exhibited lapses like missing details or factual errors. At last, we compared the rating behavior between humans and GPT-4 and found systematic bias in using GPT evaluation.
Exploring Ordinality in Text Classification: A Comparative Study of Explicit and Implicit Techniques
PDF ↗Ordinal Classification (OC) is a widely encountered challenge in Natural Language Processing (NLP), with applications in various domains such as sentiment analysis, rating prediction, and more. Previous approaches to tackle OC have primarily focused on modifying existing or creating novel loss functions that explicitly account for the ordinal nature of labels. However, with the advent of Pre-trained Language Models (PLMs), it became possible to tackle ordinality through the implicit semantics of the labels as well. This paper provides a comprehensive theoretical and empirical examination of both these approaches. Furthermore, we also offer strategic recommendations regarding the most effective approach to adopt based on specific settings.
Medical Visual Question Answering (Med-VQA) seeks to accurately respond to queries regarding medical images, a task particularly challenging for open-ended questions. This study unveils the Multi-modal Concept Alignment Pre-training (MMCAP) approach for generative Med-VQA, leveraging a knowledge graph sourced from medical image-caption datasets and the Unified Medical Language System. MMCAP advances the fusion of visual and textual medical knowledge via a graph attention network and a transformer decoder. Additionally, it incorporates a Type Conditional Prompt in the fine-tuning phase, markedly boosting the accuracy and relevance of answers to open-ended questions. Our tests on benchmark datasets illustrate MMCAP’s superiority over existing methods, demonstrating its high efficiency in data-limited settings and effective knowledge-image alignment capability.
ReactXT: Understanding Molecular “Reaction-ship” via Reaction-Contextualized Molecule-Text Pretraining
PDF ↗Molecule-text modeling, which aims to facilitate molecule-relevant tasks with a textual interface and textual knowledge, is an emerging research direction. Beyond single molecules, studying reaction-text modeling holds promise for helping the synthesis of new materials and drugs. However, previous works mostly neglect reaction-text modeling: they primarily focus on modeling individual molecule-text pairs or learning chemical reactions without texts in context. Additionally, one key task of reaction-text modeling – experimental procedure prediction – is less explored due to the absence of an open-source dataset. The task is to predict step-by-step actions of conducting chemical experiments and is crucial to automating chemical synthesis. To resolve the challenges above, we propose a new pretraining method, ReactXT, for reaction-text modeling, and a new dataset, OpenExp, for experimental procedure prediction. Specifically, ReactXT features three types of input contexts to incrementally pretrain LMs. Each of the three input contexts corresponds to a pretraining task to improve the text-based understanding of either reactions or single molecules. ReactXT demonstrates consistent improvements in experimental procedure prediction and molecule captioning and offers competitive results in retrosynthesis. Our code is available at https://github.com/syr-cn/ReactXT.
Backdoor attacks pose an increasingly severe security threat to Deep Neural Networks (DNNs) during their development stage. In response, backdoor sample purification has emerged as a promising defense mechanism, aiming to eliminate backdoor triggers while preserving the integrity of the clean content in the samples. However, existing approaches have been predominantly focused on the word space, which are ineffective against feature-space triggers and significantly impair performance on clean data. To address this, we introduce a universal backdoor defense that purifies backdoor samples in the activation space by drawing abnormal activations towards optimized minimum clean activation distribution intervals. The advantages of our approach are twofold: (1) By operating in the activation space, our method captures from surface-level information like words to higher-level semantic concepts such as syntax, thus counteracting diverse triggers; (2) the fine-grained continuous nature of the activation space allows for more precise preservation of clean content while removing triggers. Furthermore, we propose a detection module based on statistical information of abnormal activations, to achieve a better trade-off between clean accuracy and defending performance. Extensive experiments on diverse datasets and against diverse attacks (including syntax and style attacks) demonstrate that our defense achieves state-of-the-art performance.
Which Information Matters? Dissecting Human-written Multi-document Summaries with Partial Information Decomposition
PDF ↗Understanding the nature of high-quality summaries is crucial to further improve the performance of multi-document summarization. We propose an approach to characterize human-written summaries using partial information decomposition, which decomposes the mutual information provided by all source documents into union, redundancy, synergy, and unique information. Our empirical analysis on different MDS datasets shows that there is a direct dependency between the number of sources and their contribution to the summary.