General-Purpose Language Models have changed the world of Natural Language Processing, if not the world itself. The evaluation of such versatile models, while supposedly similar to evaluation of generation models before them, in fact presents a host of new evaluation challenges and opportunities. In this Tutorial, we will start from the building blocks of evaluation. The tutorial welcomes people from diverse backgrounds and assumes little familiarity with metrics, datasets, prompts and benchmarks. It will lay the foundations and explain the basics and their importance, while touching on the major points and breakthroughs of the recent era of evaluation. It will also compare traditional evaluation methods – which are still widely used – to newly developed methods. We will contrast new to old approaches, from evaluating on many-task benchmarks rather than on dedicated datasets to efficiency constraints, and from testing stability and prompts on in-context learning to using the models themselves as evaluation metrics. Finally, the tutorial will cover practical issues, ranging from reviewing widely-used benchmarks and prompt banks to efficient evaluation.

The increasing versatility of language models (LMs) has given rise to a new class of benchmarks that comprehensively assess a broad range of capabilities. Such benchmarks are associated with massive computational costs, extending to thousands of GPU hours per model. However, the efficiency aspect of these evaluation efforts had raised little discussion in the literature.In this work, we present the problem of Efficient Benchmarking, namely, intelligently reducing the computation costs of LM evaluation without compromising reliability. Using the HELM benchmark as a test case, we investigate how different benchmark design choices affect the computation-reliability trade-off. We propose to evaluate the reliability of such decisions, by using a new measure – Decision Impact on Reliability, DIoR for short.We find, for example, that a benchmark leader may change by merely removing a low-ranked model from the benchmark, and observe that a correct benchmark ranking can be obtained by considering only a fraction of the evaluation examples.Based on our findings, we outline a set of concrete recommendations for efficient benchmark design and utilization practices. To take a step further, we use our findings to propose an evaluation algorithm, that, when applied to the HELM benchmark, leads to dramatic cost savings with minimal loss of benchmark reliability, often reducing computation by x100 or more.

In the dynamic landscape of generative NLP, traditional text processing pipelines limit research flexibility and reproducibility, as they are tailored to specific dataset, task, and model combinations. The escalating complexity, involving system prompts, model-specific formats, instructions, and more, calls for a shift to a structured, modular, and customizable solution.Addressing this need, we present Unitxt, an innovative library for customizable textual data preparation and evaluation tailored to generative language models. Unitxt natively integrates with common libraries like HuggingFace and LM-eval-harness and deconstructs processing flows into modular components, enabling easy customization and sharing between practitioners. These components encompass model-specific formats, task prompts, and many other comprehensive dataset processing definitions. The Unitxt Catalog centralizes these components, fostering collaboration and exploration in modern textual data workflows. Beyond being a tool, Unitxt is a community-driven platform, empowering users to build, share, and advance their pipelines collaboratively. Join the Unitxt community at https://github.com/IBM/unitxt

Applying language models to natural language processing tasks typically relies on the representations in the final model layer, as intermediate hidden layer representations are presumed to be less informative. In this work, we argue that due to the gradual improvement across model layers, additional information can be gleaned from the contrast between higher and lower layers during inference. Specifically, in choosing between the probable next token predictions of a generative model, the predictions of lower layers can be used to highlight which candidates are best avoided. We propose a novel approach that utilizes the contrast between layers to improve text generation outputs, and show that it mitigates degenerative behaviors of the model in open-ended generation, significantly improving the quality of generated texts. Furthermore, our results indicate that contrasting between model layers at inference time can yield substantial benefits to certain aspects of general language model capabilities, more effectively extracting knowledge during inference from a given set of model parameters.

The field of Natural Language Generation (NLG) suffers from a severe shortage of labeled data due to the extremely expensive and time-consuming process involved in manual annotation. A natural approach for coping with this problem is active learning (AL), a well-known machine learning technique for improving annotation efficiency by selectively choosing the most informative examples to label. However, while AL has been well-researched in the context of text classification, its application to NLG remains largely unexplored. In this paper, we present a first systematic study of active learning for NLG, considering a diverse set of tasks and multiple leading selection strategies, and harnessing a strong instruction-tuned model. Our results indicate that the performance of existing AL strategies is inconsistent, surpassing the baseline of random example selection in some cases but not in others. We highlight some notable differences between the classification and generation scenarios, and analyze the selection behaviors of existing AL strategies. Our findings motivate exploring novel approaches for applying AL to generation tasks.

In real-world scenarios, a text classification task often begins with a cold start, when labeled data is scarce. In such cases, the common practice of fine-tuning pre-trained models, such as BERT, for a target classification task, is prone to produce poor performance. We suggest a method to boost the performance of such models by adding an intermediate unsupervised classification task, between the pre-training and fine-tuning phases. As such an intermediate task, we perform clustering and train the pre-trained model on predicting the cluster labels. We test this hypothesis on various data sets, and show that this additional classification phase can significantly improve performance, mainly for topical classification tasks, when the number of labeled instances available for fine-tuning is only a couple of dozen to a few hundred.

Recent advances in large pretrained language models have increased attention to zero-shot text classification. In particular, models finetuned on natural language inference datasets have been widely adopted as zero-shot classifiers due to their promising results and off-the-shelf availability. However, the fact that such models are unfamiliar with the target task can lead to instability and performance issues. We propose a plug-and-play method to bridge this gap using a simple self-training approach, requiring only the class names along with an unlabeled dataset, and without the need for domain expertise or trial and error. We show that fine-tuning the zero-shot classifier on its most confident predictions leads to significant performance gains across a wide range of text classification tasks, presumably since self-training adapts the zero-shot model to the task at hand.

Label Sleuth is an open source platform for building text classifiers which does not require coding skills nor machine learning knowledge.- Project website: [https://www.label-sleuth.org/](https://www.label-sleuth.org/)- Link to screencast video: [https://vimeo.com/735675461](https://vimeo.com/735675461)### AbstractText classification can be useful in many real-world scenarios, saving a lot of time for end users. However, building a classifier generally requires coding skills and ML knowledge, which poses a significant barrier for many potential users. To lift this barrier we introduce *Label Sleuth*, a free open source system for labeling and creating text classifiers. This system is unique for: - being a no-code system, making NLP accessible for non-experts. - guiding its users throughout the entire labeling process until they obtain their desired classifier, making the process efficient - from cold start to a classifier in a few hours. - being open for configuration and extension by developers. By open sourcing Label Sleuth we hope to build a community of users and developers that will widen the utilization of NLP models.

Real world scenarios present a challenge for text classification, since labels are usually expensive and the data is often characterized by class imbalance. Active Learning (AL) is a ubiquitous paradigm to cope with data scarcity. Recently, pre-trained NLP models, and BERT in particular, are receiving massive attention due to their outstanding performance in various NLP tasks. However, the use of AL with deep pre-trained models has so far received little consideration. Here, we present a large-scale empirical study on active learning techniques for BERT-based classification, addressing a diverse set of AL strategies and datasets. We focus on practical scenarios of binary text classification, where the annotation budget is very small, and the data is often skewed. Our results demonstrate that AL can boost BERT performance, especially in the most realistic scenario in which the initial set of labeled examples is created using keyword-based queries, resulting in a biased sample of the minority class. We release our research framework, aiming to facilitate future research along the lines explored here.

In Natural Language Understanding, the task of response generation is usually focused on responses to short texts, such as tweets or a turn in a dialog. Here we present a novel task of producing a critical response to a long argumentative text, and suggest a method based on general rebuttal arguments to address it. We do this in the context of the recently-suggested task of listening comprehension over argumentative content: given a speech on some specified topic, and a list of relevant arguments, the goal is to determine which of the arguments appear in the speech. The general rebuttals we describe here (in English) overcome the need for topic-specific arguments to be provided, by proving to be applicable for a large set of topics. This allows creating responses beyond the scope of topics for which specific arguments are available. All data collected during this work is freely available for research.

Extraction of financial and economic events from text has previously been done mostly using rule-based methods, with more recent works employing machine learning techniques. This work is in line with this latter approach, leveraging relevant Wikipedia sections to extract weak labels for sentences describing economic events. Whereas previous weakly supervised approaches required a knowledge-base of such events, or corresponding financial figures, our approach requires no such additional data, and can be employed to extract economic events related to companies which are not even mentioned in the training data.

Competitive debaters often find themselves facing a challenging task – how to debate a topic they know very little about, with only minutes to prepare, and without access to books or the Internet? What they often do is rely on ”first principles”, commonplace arguments which are relevant to many topics, and which they have refined in past debates. In this work we aim to explicitly define a taxonomy of such principled recurring arguments, and, given a controversial topic, to automatically identify which of these arguments are relevant to the topic. As far as we know, this is the first time that this approach to argument invention is formalized and made explicit in the context of NLP. The main goal of this work is to show that it is possible to define such a taxonomy. While the taxonomy suggested here should be thought of as a ”first attempt” it is nonetheless coherent, covers well the relevant topics and coincides with what professional debaters actually argue in their speeches, and facilitates automatic argument invention for new topics.