Morocco

Hallucinations generated by Large Language Models (LLMs) pose significant challenges for their application to low-resource languages. We present Multi-Hall-SA, a cross-lingual benchmark for hallucination detection spanning English and four low-resource South African languages: isiZulu, isiXhosa, Sepedi, and Sesotho. Derived from government texts, this benchmark categorizes hallucinations into four types aligned with established taxonomies of factual errors: temporal shifts, entity errors, numerical inaccuracies, and location mistakes. Human validation confirms the quality and cross-lingual alignment of our synthetically generated hallucinations. Our cross-lingual alignment methodology enables direct performance comparison between high-resource and low-resource languages, revealing notable gaps in detection capabilities. Evaluation across four state-of-the-art models shows they detect up to 23.6% fewer hallucinations in South African languages compared to English. Knowledge augmentation reduces this disparity, decreasing cross-lingual performance gaps by 59.4% on average. Beyond introducing a validated resource for low-resource languages, Multi-Hall-SA provides a framework for evaluating and improving factual reliability across linguistic boundaries, advancing more inclusive and equitable AI development.

EACL 2026

Multi-Hall-SA: A Cross-lingual Benchmark for Multi-Type Hallucination Detection in Low-Resource South African Languages

cross-lingual hallucination detection

low-resource african languages

knowledge augmentation

### Welcome to ACL 2023, the 61st Annual Meeting of the Association for Computational Linguistics! 
 The conference will be held in Toronto, Canada, July 9-14, 2023. 
Following the succession of the recent conferences in our field, ACL 2023 will adopt a hybrid format.
While the impact of Covid has considerably diminished in terms of traveling, obtaining visas to Canada
entails a very long process. Moreover, the global economic conditions pose challenges for many individuals to travel to conferences. Recognizing these circumstances, we know many participants may not be
able to attend the conference in person. Therefore, we are committed to providing a great virtual platform
so everyone has the opportunity to interact with other participants and enjoy the conference. Based on the
current registered participants, approxiately 30% have chosen to attend the conference virtually. Whether
you join us in person or virtually, we sincerely hope everyone has a remarkable conference experience. 
This General Chair’s message is where I express my gratitude to the many individuals who have made
enormous contributions to the conference over the past year.

Read [**ACL 2023 General Chair's message**](https://docs.google.com/document/d/1WobYM7norbG4dI48s75HfJoD89qgX5a_F-6U8AteLSA/edit?usp=sharing/) in full.

##### **[Conference Handbook](https://2023.aclweb.org/downloads/acl2023-handbook.pdf)**

ACL 2023

The Association for Computational Linguistics (ACL) is the premier international scientific and professional society for people working on computational problems involving human language, a field often referred to as either computational linguistics or natural language processing.

technical paper

#### *Message from the General Chair, Aline Villavicencio*
I’m delighted and honoured to welcome you to the 19th Conference of the European Chapter of the Association for Computational Linguistics (EACL 2026), taking place in the beautiful city of Rabat, in Morocco, in March 24-29, 2026. EACL is the flagship European conference of the Association and EACL 2026 proudly continues our field’s tradition of excellence in scholarship, innovation, and inclusivity. I am deeply grateful to the many volunteers whose dedication, generosity, and tireless efforts have made this conference possible.
For the first time EACL is being hosted in the African continent. This is an important milestone for our community, and we are grateful to our Moroccan hosts for enabling this historic moment by bringing this edition of EACL to Rabat. We are also delighted that the Second Arabic NLP School is co-located with EACL. We hope attendees enjoy this wonderful opportunity to strengthen ties with the Computational Linguistics communities across the African continent. *[Read full message](https://drive.google.com/file/d/14NlmHvwM6fPJuMmOvVh7K0vtQbEyv3SZ/view?usp=sharing)* 

<html><button style="display: inline-flex; align-items: center; justify-content: center; white-space: nowrap; border-radius: 9999px; font-weight: bold; background: #7c3aed; color: white; font-family: 'Space Grotesk', sans-serif; height: 40px; font-size: 16px; padding: 0 20px; border: none; cursor: pointer" onclick="window.open('https://underline.io/events/522/reception','_blank')">Go to Workshops and Tutorials Program</button></html>

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 EACL 2026". 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 are required to register.
Please complete your registration to continue.

We recommend reading [**the registration information**](https://2026.eacl.org/registration/) first.

**Online Registration Form**: https://acl.swoogo.com/eacl2026

Registration Required

Welcome to the 19th Conference of the European Chapter of the Association for Computational Linguistics (EACL). EACL 2026 will be held in Rabat, Morocco, from March 24–29, 2026. 

We present **MMUIE**, a large-scale universal dataset for multi-domain, document-level information extraction (IE) from long texts. Existing IE systems predominantly operate at the sentence level or within narrow domains due to annotation constraints. MMUIE addresses this gap by introducing an automated annotation pipeline that integrates traditional knowledge bases with large language models to extract fine-grained entities, aliases, and relation triples across 34 domains. The dataset comprises a weakly-supervised training set and a manually verified test set, featuring 723 entity types and 456 relation types. Empirical evaluations reveal that existing sentence-level IE models and even advanced LLMs underperform on this task, highlighting the need for better domain-aware document-level models. To this end, we develop DocUIE, a universal IE model fine-tuned on MMUIE, which achieves strong generalization and transferability across domains. MMUIE lays the foundation for robust, scalable, and universal information extraction from long-form text in diverse real-world scenarios. All code and data will be made publicly available.

MMUIE: Massive Multi-Domain Universal Information Extraction for Long Documents

Research on developmentally plausible language models has so far centered on English, leaving open questions about multilingual settings. We present a systematic study of compact models by extending BabyBERTa to EnglishвЂ“French scenarios under strictly size-matched data conditions, addressing {\it monolingual}, {\it bilingual}, and {\it cross-lingual} settings. Our design contrasts two corpus types: (i) child-directed speech (2.5M tokens), following BabyBERTa and related work, and (ii) multi-domain corpora (10M tokens), extending the BabyLM framework to French. To support fair evaluation, we also introduce new resources: French versions of QAMR and QASRL, and an English and French multi-domain corpus. We evaluate the models on both syntax and semantics tasks, comparing with Wikipedia-only training. Results reveal context-dependent effects. Training on Wikipedia consistently favors semantic tasks, while child-directed speech improves grammatical judgments in monolingual settings. Bilingual pretraining yields notable gains for Textual Entailment, with French benefiting from curriculum learning. Importantly, the same relative patterns emerge across BabyBERTa, RoBERTa, and LTG-BERT, suggesting architecture-independent findings.

Learning from Child-directed Speech in Two-language Scenarios: A French-English Case-Study

Recent attempts to leverage large language models (LLMs) for reasoning and pre-trained knowledge in multi-modal reasoning focus on two main approaches: aligning image features with linguistic space, and converting images into textual cues to exploit the implicit reasoning capabilities of LLMs. Although they integrate visual information into the reasoning pipeline, they often treat visual perception and language reasoning as separate processes, limiting the potential for fully unified multi-modal reasoning. In this paper, we propose a novel method, VisualвЂ“Linguistic Abductive Reasoning (ViLA), inspired by human abductive reasoning processes. ViLA hypothesizes a plausible answer, generates the corresponding visual and textual premises, and employs fuzzy scoring to select the most coherent combination, thus deriving the final inference. This process integrates visual and linguistic modalities into interpretable abductive reasoning chains, enabling unified multi-modal reasoning. Without fine-tuning LLMs or retrieving external knowledge, ViLA improves performance by 2.31% on AOKVQA, 1.7% on OKVQA, and 1.7% on GQA over previous state-of-the-art models, while also improving interpretability and stability.

Visual–Linguistic Abductive Reasoning with LLMs for Knowledge-based Visual Question Answering

Training complexity often scales with the size of hyperparameter space for Large Language Models (LLMs). While Direct Preference Optimization (DPO) offers learning stability through reparameterizing the reward function, its regularization against the reference policy can lead to suboptimal outcomes when the reference policy is not optimal. Recent DPO variants address this concern, but at a cost: they introduce additional hyperparameters, reducing feasibility for LLM fine-tuning. To overcome this challenge, we introduce Implicit policy Regularized Preference Optimization (IRPO), which tackles suboptimality while maintaining training simplicity. By treating the winning policy that generated the chosen responses in a pairwise dataset as an implicit policy, IRPO maximizes KL-regularized reward without extra hyperparameters. Then we propose a novel PO algorithm that directly optimizes the IRPO objective by estimating the likelihood ratio between implicit policies. As the winning policy generally outperforms the reference policy, IRPO can effectively address suboptimality. Our experiments show that IRPO significantly outperforms baseline algorithms with the same hyperparameter complexity. Moreover, IRPO demonstrates comparable performance to recent algorithms that rely on a larger number of hyperparameters, offering a practical solution for scalable LLM fine-tuning.

IRPO: Implicit Policy Regularized Preference Optimization

The rise of large language models (LLMs) has enabled us to seek answers to inherently debatable questions on LLM chatbots, necessitating a reliable way to evaluate their ability. However, traditional QA benchmarks assume fixed answers are inadequate for this purpose. To address this, we introduce DebateQA, a dataset of 2,941 debatable questions, each accompanied by multiple human-annotated partial answers that capture a variety of perspectives. We develop two metrics: Perspective Diversity, which evaluates the comprehensiveness of perspectives, and Dispute Awareness, which assesses if the LLM acknowledges the question's debatable nature. Experiments demonstrate that both metrics are aligned with human preferences and stable across different underlying models. Using DebateQA with two metrics, we assess 12 prevalent LLMs and retrieval-augmented generation methods. Our findings reveal that while LLMs generally excel at recognizing debatable issues, their ability to provide comprehensive answers encompassing diverse perspectives varies considerably.

DebateQA: Evaluating Question Answering on Debatable Knowledge

The advancement of reasoning capabilities in Large Language Models (LLMs) requires substantial amounts of high-quality reasoning data, particularly in mathematics. Existing data synthesis methods, such as data augmentation from annotated training sets or direct question generation based on relevant knowledge points and documents, have expanded datasets but face challenges in mastering the internal logic of the problem during generation and ensuring the verifiability of the solutions. To address these issues, we propose \textbf{RV-Syn}, a novel \underline{\textbf{R}}ational and \underline{\textbf{V}}erifiable mathematical \underline{\textbf{Syn}}thesis approach. RV-Syn first constructs a structured library of mathematical operations and then composes them into executable computational graphs, which serve as verifiable solution blueprints. These graphs are subsequently back-translated into complex problems, enabling solution-guided, logic-aware problem generation while inherently ensuring the verifiability of the solving process. Experimental results show RV-Syn surpasses existing synthesis methods, including those involving human-crafted problems. Our method achieves a 6.3% performance gain over the previous state-of-the-art synthetic data on LLaMA-3-8B and demonstrates superior data efficiency, outperforming others with only half the training data (50k vs. 100k), enabling a more scalable and robust reasoning dataset generation framework.

RV-Syn: Rational and Verifiable Mathematical Reasoning Data Synthesis based on Structured Function Library

Reasoning language models have set state-of-the-art (SOTA) records on many challenging benchmarks, enabled by multi-step reasoning induced by reinforcement learning. However, reasoning models are prone to generating confident, plausible responses that are incorrect (hallucinations). Knowing when and how much to trust these models is critical for safe deployment in real-world applications. To this end, we explore uncertainty quantification (UQ) of reasoning models in this work. We ask three fundamental questions: First, are reasoning models well-calibrated? Second, does deeper reasoning improve model calibration? Finally, inspired by humans' innate ability to double-check their thought processes to verify the validity of their answers and their confidence, we ask: can reasoning models improve their calibration by explicitly reasoning about their chain-of-thought traces? We introduce introspective uncertainty quantification (IUQ) to explore this direction. In extensive evaluations on SOTA reasoning models across a broad range of benchmarks, we find that reasoning models: (i) are typically overconfident, (ii) become even more overconfident with deeper reasoning, and (iii) can become better calibrated through introspection (e.g., o3-Mini and DeepSeek R1) but not uniformly (e.g., Claude 3.7 Sonnet becomes more poorly calibrated). We conclude with important research directions to design necessary UQ benchmarks and improve the calibration of reasoning models.

Reasoning about Uncertainty: Do Reasoning Models Know When They Don't Know?

Multimodal reasoning tasks such as visual question answering (VQA) require models to process both language and visual inputs. However, existing approaches typically decompose only language queries, treating images as monolithic inputs. We introduce REDI, a framework that jointly decomposes both images and questions into visual sub-domains (segmentation, material, depth, and color) with corresponding sub-questions. REDI employs an orchestrator agent that coordinates specialized worker agents, where each worker specializes in a specific visual sub-domain, and aggregates their responses through meta-reasoning. This hierarchical multi-agent design mitigates error propagation and improves compositional reasoning across both open- and closed-source MLLMs. On SEEDBench, MMBench, and CLEVR, REDI achieves absolute accuracy improvements of 8.9%, 8.2%, and 16.0%, respectively, marking a significant advancement in compositional multimodal reasoning. The code will be released on acceptance.

Compositional Reasoning via Joint Image and Language Decomposition

Identifying legal statutes given a situation is one of the most fundamental tasks in Legal NLP, and traditionally has been modeled using facts from court judgments as input, due to their abundance. However, in practical settings, the input queries are likely to be informal and asked by laymen, or non-professionals. This has led to the development of some layman LSI datasets, although there has been little research to explore the differences between court and layman data for LSI. In this work, we create ILSIC, a corpus of layman queries covering 500+ statutes from Indian law. Additionally, the corpus also contains court cases to enable researchers to effectively compare between court and layman data for LSI. We conducted extensive experiments on our corpus, including benchmarking over the layman dataset using zero and few-shot inference, retrieval-augmented generation and supervised fine-tuning. When comparing the different types of data, we observe that models trained purely on court queries are ineffective during test time on layman queries, while transfer learning from court to layman data can be beneficial in certain scenarios. We also conducted fine-grained analyses of our results in terms of categories of queries and frequency of statutes.

ILSIC: Corpora for Identifying Indian Legal Statutes from Queries by Laymen

Large language models (LLMs) are increasingly used as evaluators for natural language generation, applying human-defined rubrics to assess system outputs. However, human rubrics are often static and misaligned with how models internally represent language quality. We introduce GER-Eval (Generating Evaluation Rubrics for Evaluation) to investigate whether LLMs can design and use their own evaluation rubrics. We evaluate the semantic coherence and scoring reliability of LLM-defined criteria and their alignment with human criteria. LLMs reliably generate interpretable and task-aware evaluation dimensions and apply them within models, but their scoring reliability degrades in factual and knowledge-intensive settings. Closed-source models such as GPT-4o achieve higher agreement and cross-model generalization than open-weight models such as Llama. Our findings position evaluation as a learned linguistic capability of LLMsвЂ”consistent within models but fragmented across themвЂ”and call for new methods that jointly model human and LLM evaluative language to improve reliability and interpretability.

Downloads

Next from EACL 2026

MMUIE: Massive Multi-Domain Universal Information Extraction for Long Documents

Similar lecture

Cross-Lingual Retrieval Augmented Prompt for Low-Resource Languages