Accurate evaluation is central to the large language model (LLM) ecosystem, guiding model selection and downstream adoption across diverse use cases. In practice, however, evaluating generative outputs typically relies on rigid lexical methods to extract and assess answers, which can conflate a model's true problem-solving ability with its compliance with predefined formatting guidelines. While recent LLM-as-a-Judge approaches mitigate this issue by assessing semantic correctness rather than strict structural conformity, they also introduce substantial computational overhead, making evaluation costly. In this work, we first systematically investigate the limitations of lexical evaluation through a large-scale empirical study spanning 36 models and 15 downstream tasks, demonstrating that such methods correlate poorly with human judgments. To address this limitation, we introduce BERT-as-a-Judge, an encoder-driven approach for assessing answer correctness in reference-based generative settings, robust to variations in output phrasing, and requiring only lightweight training on synthetically annotated question-candidate-reference triplets. We show that it consistently outperforms the lexical baseline while matching the performance of much larger LLM judges, providing a compelling trade-off between the two and enabling reliable, scalable evaluation. Finally, through extensive experimentation, we provide detailed insights into BERT-as-a-Judge's performance to offer practical guidance for practitioners, and release all project artifacts to foster downstream adoption.
Stanford University（斯坦福大学）；University of California, Irvine（加州大学欧文分校）；The University of Texas at Austin（得克萨斯大学奥斯汀分校）
Because meaning can often be inferred from lexical semantics alone, word order is often a redundant cue in natural language. For example, the words chopped, chef, and onion are more likely used to convey "The chef chopped the onion," not "The onion chopped the chef." Recent work has shown large language models to be surprisingly word order invariant, but crucially has largely considered natural prototypical inputs, where compositional meaning mostly matches lexical expectations. To overcome this confound, we probe grammatical role representation in English BERT and GPT-2, on instances where le…
Contextualized word embeddings have demonstrated state-of-the-art performance in various natural language processing tasks including those that concern historical semantic change. However, language models such as BERT was trained primarily on contemporary corpus data. To investigate whether training on historical corpus data improves diachronic semantic analysis, we present a pre-trained BERT-based language model, HistBERT, trained on the balanced Corpus of Historical American English. We examine the effectiveness of our approach by comparing the performance of the original BERT and that of Hi…
Université Paris-Saclay（巴黎-萨克雷大学）；CNRS, LISN（法国国家科学研究中心，LISN）；University of Helsinki（赫尔辛基大学）
Pre-trained language models (LMs) encode rich information about linguistic structure but their knowledge about lexical polysemy remains unclear. We propose a novel experimental setup for analysing this knowledge in LMs specifically trained for different languages (English, French, Spanish and Greek) and in multilingual BERT. We perform our analysis on datasets carefully designed to reflect different sense distributions, and control for parameters that are highly correlated with polysemy such as frequency and grammatical category. We demonstrate that BERT-derived representations reflect words' …
For both human readers and pre-trained language models (PrLMs), lexical diversity may lead to confusion and inaccuracy when understanding the underlying semantic meanings of given sentences. By substituting complex words with simple alternatives, lexical simplification (LS) is a recognized method to reduce such lexical diversity, and therefore to improve the understandability of sentences. In this paper, we leverage LS and propose a novel approach which can effectively improve the performance of PrLMs in text classification. A rule-based simplification process is applied to a given sentence. P…
The Alan Turing Institute（艾伦·图灵研究所）；University of Warwick（华威大学）；Imperial College London（伦敦帝国理工学院）；Queen Mary University of London（伦敦玛丽女王大学）
Large pre-trained language models such as BERT have been the driving force behind recent improvements across many NLP tasks. However, BERT is only trained to predict missing words - either behind masks or in the next sentence - and has no knowledge of lexical, syntactic or semantic information beyond what it picks up through unsupervised pre-training. We propose a novel method to explicitly inject linguistic knowledge in the form of word embeddings into any layer of a pre-trained BERT. Our performance improvements on multiple semantic similarity datasets when injecting dependency-based and cou…
University of Cambridge（剑桥大学）；University of Mannheim（曼海姆大学）
The success of large pretrained language models (LMs) such as BERT and RoBERTa has sparked interest in probing their representations, in order to unveil what types of knowledge they implicitly capture. While prior research focused on morphosyntactic, semantic, and world knowledge, it remains unclear to which extent LMs also derive lexical type-level knowledge from words in context. In this work, we present a systematic empirical analysis across six typologically diverse languages and five different lexical tasks, addressing the following questions: 1) How do different lexical knowledge extract…
Korea University（高丽大学）；George Washington University（乔治华盛顿大学）
Recent pre-trained language models (PLMs) achieved great success on many natural language processing tasks through learning linguistic features and contextualized sentence representation. Since attributes captured in stacked layers of PLMs are not clearly identified, straightforward approaches such as embedding the last layer are commonly preferred to derive sentence representations from PLMs. This paper introduces the attention-based pooling strategy, which enables the model to preserve layer-wise signals captured in each layer and learn digested linguistic features for downstream tasks. The …
UNC Chapel Hill（北卡罗来纳大学教堂山分校）；Penn State University（宾夕法尼亚州立大学）
Comments, or natural language descriptions of source code, are standard practice among software developers. By communicating important aspects of the code such as functionality and usage, comments help with software project maintenance. However, when the code is modified without an accompanying correction to the comment, an inconsistency between the comment and code can arise, which opens up the possibility for developer confusion and bugs. In this paper, we propose two models based on BERT (Devlin et al., 2019) and Longformer (Beltagy et al., 2020) to detect such inconsistencies in a natural …
This paper describes a system submitted by team BigGreen to LCP 2021 for predicting the lexical complexity of English words in a given context. We assemble a feature engineering-based model with a deep neural network model founded on BERT. While BERT itself performs competitively, our feature engineering-based model helps in extreme cases, eg. separating instances of easy and neutral difficulty. Our handcrafted features comprise a breadth of lexical, semantic, syntactic, and novel phonological measures. Visualizations of BERT attention maps offer insight into potential features that Transforme…
Contextualized embeddings such as BERT can serve as strong input representations to NLP tasks, outperforming their static embeddings counterparts such as skip-gram, CBOW and GloVe. However, such embeddings are dynamic, calculated according to a sentence-level context, which limits their use in lexical semantics tasks. We address this issue by making use of dynamic embeddings as word representations in training static embeddings, thereby leveraging their strong representation power for disambiguating context information. Results show that this method leads to improvements over traditional stati…
Lexical Prompt Compression for Large Language Models: A Training-Free, Deterministic Pipeline with Empirical Pareto Analysis Across Eleven Task Categories
Institute for Natural Language Processing, University of Stuttgart（斯图加特大学自然语言处理研究所）
Type- and token-based embedding architectures are still competing in lexical semantic change detection. The recent success of type-based models in SemEval-2020 Task 1 has raised the question why the success of token-based models on a variety of other NLP tasks does not translate to our field. We investigate the influence of a range of variables on clusterings of BERT vectors and show that its low performance is largely due to orthographic information on the target word, which is encoded even in the higher layers of BERT representations. By reducing the influence of orthography we considerably …
Google Research（谷歌研究院）；New York University（纽约大学）；Brown University（布朗大学）
Pre-trained language models perform well on a variety of linguistic tasks that require symbolic reasoning, raising the question of whether such models implicitly represent abstract symbols and rules. We investigate this question using the case study of BERT's performance on English subject-verb agreement. Unlike prior work, we train multiple instances of BERT from scratch, allowing us to perform a series of controlled interventions at pre-training time. We show that BERT often generalizes well to subject-verb pairs that never occurred in training, suggesting a degree of rule-governed behavior.…
École Normale Supérieure-PSL（巴黎高等师范学院-PSL大学）；CNRS（法国国家科学研究中心）；U. Sorbonne Nouvelle（新索邦大学）；University of Pisa（比萨大学）
Although transformer-based Neural Language Models demonstrate impressive performance on a variety of tasks, their generalization abilities are not well understood. They have been shown to perform strongly on subject-verb number agreement in a wide array of settings, suggesting that they learned to track syntactic dependencies during their training even without explicit supervision. In this paper, we examine the extent to which BERT is able to perform lexically-independent subject-verb number agreement (NA) on targeted syntactic templates. To do so, we disrupt the lexical patterns found in natu…
Purdue University（普渡大学）；University of Chicago（芝加哥大学）
Models trained to estimate word probabilities in context have become ubiquitous in natural language processing. How do these models use lexical cues in context to inform their word probabilities? To answer this question, we present a case study analyzing the pre-trained BERT model with tests informed by semantic priming. Using English lexical stimuli that show priming in humans, we find that BERT too shows "priming," predicting a word with greater probability when the context includes a related word versus an unrelated one. This effect decreases as the amount of information provided by the con…
Pretrained Masked Language Models (MLMs) have revolutionised NLP in recent years. However, previous work has indicated that off-the-shelf MLMs are not effective as universal lexical or sentence encoders without further task-specific fine-tuning on NLI, sentence similarity, or paraphrasing tasks using annotated task data. In this work, we demonstrate that it is possible to turn MLMs into effective universal lexical and sentence encoders even without any additional data and without any supervision. We propose an extremely simple, fast and effective contrastive learning technique, termed Mirror-B…
Université Paris-Saclay（巴黎-萨克雷大学）；CNRS, LIMSI（法国国家科学研究中心，LIMSI实验室）；University of Helsinki（赫尔辛基大学）
We present the MULTISEM systems submitted to SemEval 2020 Task 3: Graded Word Similarity in Context (GWSC). We experiment with injecting semantic knowledge into pre-trained BERT models through fine-tuning on lexical semantic tasks related to GWSC. We use existing semantically annotated datasets and propose to approximate similarity through automatically generated lexical substitutes in context. We participate in both GWSC subtasks and address two languages, English and Finnish. Our best English models occupy the third and fourth positions in the ranking for the two subtasks. Performance is low…
Meta AI；MIT（麻省理工学院）；Johns Hopkins（约翰斯·霍普金斯大学）；Purdue University（普渡大学）
Targeted syntactic evaluations of language models ask whether models show stable preferences for syntactically acceptable content over minimal-pair unacceptable inputs. Most targeted syntactic evaluation datasets ask models to make these judgements with just a single context-free sentence as input. This does not match language models' training regime, in which input sentences are always highly contextualized by the surrounding corpus. This mismatch raises an important question: how robust are models' syntactic judgements in different contexts? In this paper, we investigate the stability of lan…
Wee Kim Wee School of Communication and Information（黄金辉传播与信息学院）；Nanyang Technological University（南洋理工大学）
Semantic parsing is the task of transforming sentences from natural language into formal representations of predicate-argument structures. Under this research area, frame-semantic parsing has attracted much interest. This parsing approach leverages the lexical information defined in FrameNet to associate marked predicates or targets with semantic frames, thereby assigning semantic roles to sentence components based on pre-specified frame elements in FrameNet. In this paper, a deep neural network architecture known as Positional Attention-based Frame Identification with BERT (PAFIBERT) is prese…
Many NLP tasks have benefited from transferring knowledge from contextualized word embeddings, however the picture of what type of knowledge is transferred is incomplete. This paper studies the types of linguistic phenomena accounted for by language models in the context of a Conversational Question Answering (CoQA) task. We identify the problematic areas for the finetuned RoBERTa, BERT and DistilBERT models through systematic error analysis - basic arithmetic (counting phrases), compositional semantics (negation and Semantic Role Labeling), and lexical semantics (surprisal and antonymy). When…
This paper introduces DaN+, a new multi-domain corpus and annotation guidelines for Danish nested named entities (NEs) and lexical normalization to support research on cross-lingual cross-domain learning for a less-resourced language. We empirically assess three strategies to model the two-layer Named Entity Recognition (NER) task. We compare transfer capabilities from German versus in-language annotation from scratch. We examine language-specific versus multilingual BERT, and study the effect of lexical normalization on NER. Our results show that 1) the most robust strategy is multi-task lear…
Alibaba Group（阿里巴巴集团）；School of Computer Science and Technology, East China Normal University（华东师范大学计算机科学与技术学院）
Lexical relations describe how concepts are semantically related, in the form of relation triples. The accurate prediction of lexical relations between concepts is challenging, due to the sparsity of patterns indicating the existence of such relations. We propose the Knowledge-Enriched Meta-Learning (KEML) framework to address the task of lexical relation classification. In KEML, the LKB-BERT (Lexical Knowledge Base-BERT) model is presented to learn concept representations from massive text corpora, with rich lexical knowledge injected by distant supervision. A probabilistic distribution of au…
We present the results of our participation in the DIACR-Ita shared task on lexical semantic change detection for Italian. We exploit Average Pairwise Distance of token-based BERT embeddings between time points and rank 5 (of 8) in the official ranking with an accuracy of $.72$. While we tune parameters on the English data set of SemEval-2020 Task 1 and reach high performance, this does not translate to the Italian DIACR-Ita data set. Our results show that we do not manage to find robust ways to exploit BERT embeddings in lexical semantic change detection.
Hefei University of Technology（合肥工业大学）；Mininglamp Academy of Sciences（明略科学院）
Lexical simplification (LS) aims to replace complex words in a given sentence with their simpler alternatives of equivalent meaning, to simplify the sentence. Recently unsupervised lexical simplification approaches only rely on the complex word itself regardless of the given sentence to generate candidate substitutions, which will inevitably produce a large number of spurious candidates. In this paper, we propose a lexical simplification framework LSBert based on pretrained representation model Bert, that is capable of (1) making use of the wider context when both detecting the words in need o…
University of Massachusetts, Amherst（马萨诸塞大学阿默斯特分校）
Phrase representations derived from BERT often do not exhibit complex phrasal compositionality, as the model relies instead on lexical similarity to determine semantic relatedness. In this paper, we propose a contrastive fine-tuning objective that enables BERT to produce more powerful phrase embeddings. Our approach (Phrase-BERT) relies on a dataset of diverse phrasal paraphrases, which is automatically generated using a paraphrase generation model, as well as a large-scale dataset of phrases in context mined from the Books3 corpus. Phrase-BERT outperforms baselines across a variety of phrase-…
Bosch Center for Artificial Intelligence（博世人工智能中心）；Carnegie Mellon University（卡内基梅隆大学）
We study the utility of the lexical translation model (IBM Model 1) for English text retrieval, in particular, its neural variants that are trained end-to-end. We use the neural Model1 as an aggregator layer applied to context-free or contextualized query/document embeddings. This new approach to design a neural ranking system has benefits for effectiveness, efficiency, and interpretability. Specifically, we show that adding an interpretable neural Model 1 layer on top of BERT-based contextualized embeddings (1) does not decrease accuracy and/or efficiency; and (2) may overcome the limitation …
For many (minority) languages, the resources needed to train large models are not available. We investigate the performance of zero-shot transfer learning with as little data as possible, and the influence of language similarity in this process. We retrain the lexical layers of four BERT-based models using data from two low-resource target language varieties, while the Transformer layers are independently fine-tuned on a POS-tagging task in the model's source language. By combining the new lexical layers and fine-tuned Transformer layers, we achieve high task performance for both target langua…
Huawei Noah’s Ark Lab（华为诺亚方舟实验室）；The Hong Kong University of Science and Technology（香港科技大学）；Huawei Technologies Co., Ltd.（华为技术有限公司）
Incorporating lexical knowledge into deep learning models has been proved to be very effective for sequence labeling tasks. However, previous works commonly have difficulty dealing with large-scale dynamic lexicons which often cause excessive matching noise and problems of frequent updates. In this paper, we propose DyLex, a plug-in lexicon incorporation approach for BERT based sequence labeling tasks. Instead of leveraging embeddings of words in the lexicon as in conventional methods, we adopt word-agnostic tag embeddings to avoid re-training the representation while updating the lexicon. Mor…
University of Leeds（利兹大学）；Alexandria University（亚历山大大学）
This paper presents an attempt to build a Modern Standard Arabic (MSA) sentence-level simplification system. We experimented with sentence simplification using two approaches: (i) a classification approach leading to lexical simplification pipelines which use Arabic-BERT, a pre-trained contextualised model, as well as a model of fastText word embeddings; and (ii) a generative approach, a Seq2Seq technique by applying a multilingual Text-to-Text Transfer Transformer mT5. We developed our training corpus by aligning the original and simplified sentences from the internationally acclaimed Arabic …
