Austria

Large language models (LLMs) often demonstrate strong performance by leveraging implicit knowledge acquired during pretraining. Analogical reasoning, which solves new problems by referencing similar known examples, offers a structured way to utilize this knowledge, but can also lead to subtle factual errors and hallucinations. In this work, we investigate whether LLMs can recognize the reliability of their own analogical outputs using black-box uncertainty estimation (UE). We evaluate six UE metrics across two reasoning-intensive tasks: mathematical problem solving (GSM8K) and code generation (Codeforces). Our results show that Kernel Language Entropy (KLE) and Lexical Similarity (LexSim) are the most robust indicators of correctness. Moreover, while analogical prompting increases model confidence over direct prompting, most uncertainty arises during the analogy transfer step. These findings highlight the limitations of analogical knowledge transfer in LLMs and demonstrate the potential of UE methods for detecting hallucinated reasoning in black-box settings.

ACL 2025

Can LLMs Recognize Their Own Analogical Hallucinations? Evaluating Uncertainty Estimation for Analogical Reasoning

**Welcome to EMNLP 2023!**

On behalf of the EMNLP 2023 Organizing Committee, I extend a warm and heartfelt welcome to all of you to the 2023 Conference on Empirical Methods in Natural Language Processing (EMNLP). It is with immense pleasure and excitement that we gather here in Singapore, a vibrant hub of innovation and technological advancement.

The conference program is packed with insightful presentations, thought-provoking workshops, and engaging networking opportunities. In addition to the technical sessions, we have also planned several social events that will provide you with the opportunity to connect with your colleagues.

The conference is held in person at the Resorts World Convention Centre in Singapore, and available on-line with the help of Underline.

For those in Singapore, I hope that you will find time to explore the exciting Sentosa Island, Gardens by the Bay, Singapore Botanic Gardens and the many other attractions unique to Singapore.

Of course, we are grateful to our sponsors and partners for their generous support of EMNLP 2023, whose contributions make it possible for us to host this world-class event.

Yuji Matsumoto (RIKEN AIP) 
EMNLP 2023 General Chair

To access the **EMNLP 2023** event page on Underline, you need to register for the Conference. 
Please follow **[this link](https://2023.emnlp.org/registration/)** for more details.

EMNLP 2023

EMNLP 2023 took place in Singapore from Dec 6th to Dec 10th, 2023.

workshop paper

### Welcome to The 63rd Annual Meeting of the Association for Computational Linguistics (ACL 2025)

Message from the General Chair: 
*It is my great pleasure and honor to welcome you to the 63rd Annual Meeting of the Association for Computational Linguistics (ACL 2025), held in beautiful Vienna, Austria, from July 27 to August 1, 2025. ACL2025continues our field’s tradition of excellence in scholarship, innovation, and inclusivity, and I am deeply grateful to the many volunteers who have worked tirelessly to bring this event to life.* 
[Read more](https://drive.google.com/file/d/1GI_hvOpjswAuYdUTromfeDiPpCcqidwg/view?usp=sharing)

To access this event page, you need to log in with the **email address you registered with**. Access credentials will be sent to your email from Underline - subject line "Welcome to ACL 2025". Please be sure to check your spam email folder if you do not see an email confirmation right away.

Please log in

To access this event page, you need to log in with the **email address you registered with**. 

Welcome to The 63rd Annual Meeting of the Association for Computational Linguistics (ACL 2025)

We present Meetalk, a retrieval-augmented and knowledge-adaptive system for generating personalized meeting minutes. Although large language models (LLMs) excel at summarizing, their output often lacks faithfulness and does not reflect user-specific structure and style. Meetalk addresses these issues by integrating ASR-based transcription with LLM generation guided by user-derived knowledge. Specifically, Meetalk maintains and updates three structured databases, Table of Contents, Chapter Allocation, and Writing Style, based on user-uploaded samples and editing feedback. These serve as a dynamic memory that is retrieved during generation to ground the model’s outputs. To further enhance reliability, Meetalk introduces hallucination-aware uncertainty markers that highlight low-confidence segments for user review. In a user study in five real-world meeting scenarios, Meetalk significantly outperforms a strong baseline (iFLYTEK ASR + ChatGPT-4o) in completeness, contextual relevance, and user trust. Our findings underscore the importance of knowledge foundation and feedback-driven adaptation in building trustworthy, personalized LLM systems for high-stakes summarization tasks.

Meetalk: Retrieval-Augmented and Adaptively Personalized Meeting Summarization with Knowledge Learning from User Corrections

Large language models (LLMs) have shown strong performance across natural language reasoning tasks, yet their reasoning processes remain brittle and difficult to interpret. Prompting techniques like Chain-of-Thought (CoT) enhance reliability by eliciting intermediate reasoning steps or aggregating multiple outputs. However, they lack mechanisms for enforcing logical structure and assessing internal coherence. We introduce Theorem-of-Thought (ToTh), a novel framework that models reasoning as collaboration among three parallel agents, each simulating a distinct mode of inference: abductive, deductive, and inductive. Each agent produces a reasoning trace, which is structured into a formal reasoning graph. To evaluate consistency, we apply Bayesian belief propagation guided by natural language inference (NLI), assigning confidence scores to each step. The most coherent graph is selected to derive the final answer. Experiments on symbolic (WebOfLies) and numerical (MultiArith) reasoning benchmarks show that ToTh consistently outperforms CoT, Self-Consistency, and CoT-Decoding across multiple LLMs, while producing interpretable and logically grounded reasoning chains. Our findings suggest a promising direction for building more robust and cognitively inspired LLM reasoning. The implementation is available at https://github.com/KurbanIntelligenceLab/theorem-of-thought.

Theorem-of-Thought: A Multi-Agent Framework for Abductive, Deductive, and Inductive Reasoning in Language Models

Chinese homophones, prevalent in Internet culture, bring rich linguistic twists that are challenging for language models. While native speakers disambiguate them through phonological reasoning and contextual understanding, it remains untested how well LLMs perform on this task and whether LLMs also achieve this via similar reasoning processes or merely through memorization of homophone-original word pairs during training.

In this paper, we present HomoP-CN, the first Chinese Internet homophones dataset with systematic perturbations for evaluating LLMs' homophone restoration capabilities. Using this benchmark, we investigated the influence of semantic, phonological, and graphemic features on LLMs' restoration accuracy, measured the reliance levels of each model on memorization during restoration through consistency ratios under controlled perturbations, and assessed the effectiveness of various prompting strategies, including contextual cues, pinyin augmentation, few-shot learning, and thought-chain approaches.

Reasoning or Memorization? Investigating LLMs’ Capabilit in Restoring Chinese Internet Homophones

Large language models (LLMs) achieve high performance through instruction-tuning, which involves learning various tasks using instruction templates. However, these templates often contain task-specific expressions, which are words that frequently appear in certain contexts but do not always convey the actual meaning of that context, even if they seem closely related to the target task. Biases inherent in such instruction templates may be learned by LLMs during training, potentially degrading performance when the models encounter superficial expressions. In this study, we propose a method that incorporates additional instructions to FLAN templates, without altering the base instruction to produce "superfluous instructions". This allows us to investigate the vulnerabilities of LLMs caused by overfitting to task-specific expressions embedded in instruction templates. The experimental results revealed that the inclusion of superficial words strongly related to each task in the instruction text can alter the output, regardless of the intended meaning.

Superfluous Instruction: Vulnerabilities Stemming from Task-Specific Superficial Expressions in Instruction Templates

Dense vector representations have become foundational to modern natural language processing (NLP), powering diverse workflows from semantic search and retrieval augmented generation to content comparison across languages. 
Although Wikipedia is one of the most comprehensive and widely used datasets in modern NLP research, it lacks a fully reproducible and permissively licensed dense vectorization.
In this paper, we present Wikivecs, a fully reproducible, permissively licensed dataset containing dense vector embeddings for every article in Multilingual Wikipedia. 
Our pipeline leverages a fully reproducible and permissively licensed multilingual text encoder to embed Wikipedia articles into a unified vector space, making it easy to compare and analyze content across languages.
Alongside these vectors, we release a two-dimensional data map derived from the vectors, enabling visualization and exploration of Multilingual Wikipedia’s content landscape.
We demonstrate the utility of our dataset by identifying several content gaps between English and Russian Wikipedia.

Wikivecs: A Fully Reproducible Vectorization of Multilingual Wikipedia

Specialized entity linking (EL) models are well-trained at mapping mentions to unique knowledge base (KB) entities according to a given context.
However, specialized EL models struggle to disambiguate long-tail entities due to their limited training data.
Meanwhile, extensively pre-trained large language models (LLMs) possess broader knowledge of uncommon entities.
Yet, with a lack of specialized EL training, LLMs frequently fail to generate accurate KB entity names, limiting their standalone effectiveness in EL.
With the observation that LLMs are more adept at context generation instead of EL execution, 
we introduce LLM-Augmented Entity Linking (LLMAEL), 
the first framework to enhance specialized EL models with LLM data augmentation.
LLMAEL leverages off-the-shelf, tuning-free LLMs as context augmenters, generating entity descriptions to serve as additional input for specialized EL models.
Experiments show that LLMAEL sets new state-of-the-art results across 6 widely adopted EL benchmarks: compared to prior methods that integrate tuning-free LLMs into EL, LLMAEL achieves an absolute 8.9\% gain in EL accuracy. 
We release our datasets and models to facilitate future research.

LLMaEL: Large Language Models are Good Context Augmenters for Entity Linking

Wikipedia, as a massively multilingual, community-driven platform, is a valuable resource for Natural Language Processing (NLP), yet the consistency of honorific usage in honorific-rich languages remains underexplored.
Honorifics, subtle yet profound linguistic markers, encode social hierarchies, politeness norms, and cultural values, but Wikipedia's editorial guidelines lack clear standards for their usage in languages where such forms are grammatically and socially prevalent. 
This paper addresses this gap through a large-scale, analysis of third-person honorific pronouns and verb forms in Hindi and Bengali Wikipedia articles. Using Large Language Models (LLM), we automatically annotate $10,000$ articles per language for honorific usage and socio-demographic features such as gender, age, fame, and cultural origin.
We investigate: (i) the consistency of honorific usage across articles, (ii) how inconsistencies correlate with socio-cultural factors, and (iii) the presence of explicit or implicit biases across languages.
We find that honorific usage is consistently more common in Bengali than Hindi, while non-honorific forms are more frequent for infamous, juvenile, and exotic entities in both. Notably, gender bias emerges in both languages, particularly in Hindi, where men are more likely to receive honorifics than women.
Our analysis highlights the need for Wikipedia to develop language-specific editorial guidelines for honorific usage.

Women, Infamous, and Exotic Beings: What Honorific Usages in Wikipedia Reflect on the Cross-Cultural Sociolinguistic Norms?

Given Wikipedia’s role as a trusted source of high-quality, reliable content, there are growing concerns about the proliferation of low-quality machine-generated text (MGT) produced by large language models (LLMs) on its platform. Reliable detection of MGT is therefore essential, yet existing work primarily evaluates MGT detectors on generic generation tasks, rather than on tasks more commonly performed by Wikipedia editors. This misalignment can lead to poor generalisability when applied to real-world Wikipedia contexts.
We introduce WETBench, a multilingual, multi-generator, and task-specific benchmark for MGT detection. We define three editing tasks empirically grounded in Wikipedia editors’ perceived use cases for LLM-assisted editing: Paragraph Writing, Summarisation, and Text Style Transfer, which we implement using two new datasets across three languages. For each writing task, we evaluate three prompts, produce MGT across multiple generators using the best-performing prompt, and benchmark diverse detectors.
We find that, across settings, training-based detectors achieve an average accuracy of 78%, while zero-shot detectors average 58%. These results demonstrate that detectors struggle with MGT in realistic generation scenarios and underscore the importance of evaluating such models on diverse, task-specific data to assess their reliability in editor-driven contexts.

WETBench: A Benchmark for Detecting Task-Specific Machine-Generated Text on Wikipedia

Wikipedia is powered by MediaWiki, a free and open-source software that is also the infrastructure for many other wiki-based online encyclopedias. These include the recently launched website Ruwiki, which has copied and modified the original Russian Wikipedia content to conform to Russian law. To identify practices and narratives that could be associated with different forms of knowledge manipulation, this article presents an in-depth analysis of this Russian Wikipedia fork. We propose a methodology to characterize the main changes with respect to the original version. The foundation of this study is a comprehensive comparative analysis of more than 1.9M articles from Russian Wikipedia and its fork. Using meta-information and geographical, temporal, categorical, and textual features, we explore the changes made by Ruwiki editors. Furthermore, we present a classification of the main topics of knowledge manipulation in this fork, including a numerical estimation of their scope. This research not only sheds light on significant changes within Ruwiki, but also provides a methodology that could be applied to analyze other Wikipedia forks and similar collaborative projects.

Characterizing Knowledge Manipulation in a Russian Wikipedia Fork

Proper nouns in Arabic Wikipedia are frequently undiacritized, creating ambiguity in pronunciation and interpretation, especially for transliterated named entities of foreign origin. While transliteration and diacritization have been well-studied separately in Arabic NLP, their intersection remains underexplored. In this paper, we introduce a new manually diacritized dataset of Arabic proper nouns of various origins with their English Wikipedia equivalent glosses, and present the challenges and guidelines we followed to create it. We benchmark GPT-4o on the task of recovering full diacritization given the undiacritized Arabic and English forms, and analyze its performance. Achieving 73\% accuracy, our results underscore both the difficulty of the task and the need for improved models and resources. We release our dataset to facilitate further research on Arabic Wikipedia proper noun diacritization.

Downloads

Next from ACL 2025

Meetalk: Retrieval-Augmented and Adaptively Personalized Meeting Summarization with Knowledge Learning from User Corrections

Similar lecture

Can language models learn analogical reasoning? Investigating training objectives and comparisons to human performance

Stay up to date with the latest Underline news!

PRESENTATIONS

CONFERENCES

COMPANY

RESOURCES

.css-70qvj9{display:-webkit-box;display:-webkit-flex;display:-ms-flexbox;display:flex;-webkit-align-items:center;-webkit-box-align:center;-ms-flex-align:center;align-items:center;}Downloads

Next from ACL 2025

Meetalk: Retrieval-Augmented and Adaptively Personalized Meeting Summarization with Knowledge Learning from User Corrections

Similar lecture

Can language models learn analogical reasoning? Investigating training objectives and comparisons to human performance

Stay up to date with the latest Underline news!

PRESENTATIONS

CONFERENCES

COMPANY

RESOURCES

Downloads