China

In the age of misinformation, hallucination---the tendency of Large Language Models (LLMs) to generate non-factual or unfaithful responses---represents the main risk for their global utility. Despite LLMs becoming increasingly multilingual, the vast majority of research on detecting and quantifying LLM hallucination are (a) English-centric and (b) focus on machine translation (MT) and summarization, tasks that are less common ``in the wild&#39;&#39; than open information seeking. In contrast, we aim to quantify the extent of LLM hallucination across languages in knowledge-intensive long-form question answering (LFQA). To this end, we train a multilingual hallucination detection model and conduct a large-scale study across 30 languages and 6 open-source LLM families. We start from an English hallucination detection dataset and rely on MT to translate-train a detection model. We also manually annotate gold data for five high-resource languages; we then demonstrate, for these languages, that the estimates of hallucination rates are similar between silver (LLM-generated) and gold test sets, validating the use of silver data for estimating hallucination rates for other languages. For the final rates estimation, we build open-domain QA dataset for 30 languages with LLM-generated prompts and Wikipedia articles as references. Our analysis shows that LLMs, in absolute terms, hallucinate more tokens in high-resource languages due to longer responses, but that the actual hallucination rates (i.e., normalized for length) seems uncorrelated with the sizes of languages&#39; digital footprints. We also find that smaller LLMs hallucinate more, and significantly, LLMs with broader language support display higher hallucination rates.

EMNLP 2025

How Much Do LLMs Hallucinate across Languages? On Realistic Multilingual Estimation of LLM Hallucination

llms

hallucinations

multilingual

## Welcome!
"I am excited to welcome you to this year’s edition of the Conference on Empirical Methods in Natural Language Processing! Importantly, it marks the 30th edition of EMNLP. With over 8,000 submissions, more than 3,000 accepted papers, and thousands of attendees, we have come a long way from that first
workshop, which had 14 accepted papers. As the field looks ahead, Suzhou is the fitting location for celebrating this milestone: rooted in a long literary tradition, yet modern and forward-looking, and home to a large share of the NLP community."<br>

*Message from the General Chair, Dirk Hovy*





<br>

Celebrate 30 Years of EMNLP! 
EMNLP 2025 will be held in Suzhou, China from November 5th to November 9th, 2025.

In the age of misinformation, hallucination---the tendency of Large Language Models (LLMs) to generate non-factual or unfaithful responses---represents the main risk for their global utility. Despite LLMs becoming increasingly multilingual, the vast majority of research on detecting and quantifying LLM hallucination are (a) English-centric and (b) focus on machine translation (MT) and summarization, tasks that are less common ``in the wild'' than open information seeking. In contrast, we aim to quantify the extent of LLM hallucination across languages in knowledge-intensive long-form question answering (LFQA). To this end, we train a multilingual hallucination detection model and conduct a large-scale study across 30 languages and 6 open-source LLM families. We start from an English hallucination detection dataset and rely on MT to translate-train a detection model. We also manually annotate gold data for five high-resource languages; we then demonstrate, for these languages, that the estimates of hallucination rates are similar between silver (LLM-generated) and gold test sets, validating the use of silver data for estimating hallucination rates for other languages. For the final rates estimation, we build open-domain QA dataset for 30 languages with LLM-generated prompts and Wikipedia articles as references. Our analysis shows that LLMs, in absolute terms, hallucinate more tokens in high-resource languages due to longer responses, but that the actual hallucination rates (i.e., normalized for length) seems uncorrelated with the sizes of languages' digital footprints. We also find that smaller LLMs hallucinate more, and significantly, LLMs with broader language support display higher hallucination rates.

poster

Instructing language models with user intent requires large instruction datasets, which are only available for a limited set of languages. In this paper, we explore alternatives to conventional instruction adaptation pipelines in low-resource scenarios. We assume a realistic scenario for low-resource languages, where only the following are available: corpora in the target language, existing open-weight multilingual base and instructed backbone LLMs, and synthetically generated instructions sampled from the instructed backbone. We present a comprehensive set of experiments for Basque that systematically study different combinations of these components evaluated on benchmarks and human preferences from 1680 participants. Our conclusions show that target language corpora are essential, with synthetic instructions yielding robust models, and, most importantly, that using as backbone an instruction-tuned model outperforms using a base non-instructed model, and improved results when scaling up. Using Llama 3.1 instruct 70B as backbone our model comes near frontier models of much larger sizes for Basque, without using any Basque data apart from the 1.2B word corpora. We release code, models, instruction datasets, and human preferences to support full reproducibility in future research on low-resource language adaptation.

Instructing Large Language Models for Low-Resource Languages: A Systematic Study for Basque

Large Reasoning Models (LRMs) have demonstrated impressive performances across diverse domains. However, how safety of Large Language Models (LLMs) benefits from enhanced reasoning capabilities against jailbreak queries remains unexplored. To bridge this gap, in this paper, we propose Reasoning-to-Defend (R2D), a novel training paradigm that integrates a safety-aware reasoning mechanism into LLMs' generation. This enables self-evaluation at each step of the reasoning process, forming safety pivot tokens as indicators of the safety status of responses. Furthermore, in order to improve the accuracy of predicting pivot tokens, we propose Contrastive Pivot Optimization (CPO), which enhances the model's perception of the safety status of given dialogues. LLMs dynamically adjust their response strategies during reasoning, significantly enhancing their safety capabilities defending jailbreak attacks. Extensive experiments demonstrate that R2D effectively mitigates various attacks and improves overall safety, while maintaining the original performances. This highlights the substantial potential of safety-aware reasoning in improving robustness of LRMs and LLMs against various jailbreaks.

Reasoning-to-Defend: Safety-Aware Reasoning Can Defend Large Language Models from Jailbreaking

Extremely low-resource languages, especially those written in rare scripts, remain largely unsupported by large language models (LLMs). This is due in part to compounding factors such as the lack of training data. This paper delivers the first comprehensive analysis of whether LLMs can acquire such languages purely via in-context learning (ICL), with or without auxiliary alignment signals, and how these methods compare to parameter-efficient fine-tuning (PEFT). We systematically evaluate 20 under-represented languages across three state-of-the-art multilingual LLMs. Our findings highlight the limitation of PEFT when both language and its script are extremely under-represented by the LLM. In contrast, zero-shot ICL with language alignment is impressively effective on extremely low-resource languages, while few-shot ICL or PEFT is more beneficial for languages relatively better represented by LLMs. For LLM practitioners working on extremely low-resource languages, we summarise guidelines grounded by our results on adapting LLMs to low-resource languages, e.g., avoiding fine-tuning a multilingual model on languages of unseen scripts.

It's All About In-Context Learning! Teaching Extremely Low-Resource Languages to LLMs

As Large Language Models (LLMs) become increasingly prevalent, their generated outputs are proliferating across the web, risking a future where machine-generated content dilutes human-authored text. Since online data is the primary resource for LLM pre-training, subsequent models could be trained on an unknown portion of synthetic samples. This will lead to model collapse, a degenerative process whereby LLMs reinforce their own errors, converge to a low variance output distribution, and ultimately yield a declining performance. In this study, we investigate the impact of decoding strategy on model collapse, analysing the text characteristics at each model generation, the similarity to human references, and the resulting model performance. Using the decoding strategies that lead to the most significant degradation, we evaluate model collapse in more realistic scenarios where the origin of the data (human or synthetic) is unknown. We train a machine-generated text detector and propose an importance sampling approach to alleviate model collapse. Our method is validated on two LLM variants (GPT-2 and SmolLM2), across a range of model sizes (124M to 1.7B), on the open-ended text generation task. We demonstrate that it can not only prevent model collapse but also improve performance when sufficient human-authored samples are present.

Machine-generated text detection prevents language model collapse

We propose Mixture of Languages (MoL), a new strategy to pretrain largely multilingual encoders. Recent work in this field has relied on training transformer encoders on a large amount of multilingual data, with all parameters shared across all languages, without studying how to optimally balance language transfer and interference to achieve better performance. To address this, MoL proposes to group languages based on their similarity, and add parallel, sparsely activated layers that process each group independently. This architecture allows MoL to boost language transfer while minimizing interference, without increasing the active parameter count. We show that MoL largely outperforms a dense counterpart trained with the same configuration, as well as MoE models and public multilingual encoders such as XLM-R or mBERT on downstream tasks.

Mixture of Languages: Improved Multilingual Encoders Through Language Grouping

Arabic dialects form a diverse continuum, yet NLP models often treat them as discrete categories. Recent work addresses this issue by modeling dialectness as a continuous variable, notably through the Arabic Level of Dialectness (ALDi). However, ALDi reduces complex variation to a single dimension. We propose a complementary measure: the Arabic Generality Score (AGS), which quantifies how widely a word is used across dialects. We introduce a pipeline that combines word alignment, etymology-aware edit distance, and smoothing to annotate a parallel corpus with word-level AGS. A regression model is then trained to predict AGS in context. Our approach outperforms strong baselines, including state-of-the-art dialect ID systems, on a multi-dialect benchmark. AGS offers a scalable, linguistically grounded way to model lexical generality, enriching representations of Arabic dialectness.

The Arabic Generality Score: Another Dimension of Modeling Arabic Dialectness

Despite Greece’s pivotal role in the global economy, large language models (LLMs) remain underexplored for Greek financial context due to the linguistic complexity of Greek and the scarcity of domain-specific datasets. While multilingual financial NLP has revealed large performance gaps across languages, no benchmarks or LLMs have been tailored for Greek financial tasks until now. To bridge this gap, we introduce Plutus-ben, the first Greek Financial Evaluation Benchmark, and Plutus-8B, the first financial LLM fine-tuned on Greek-specific financial data. Plutus-ben addresses five core tasks: numeric/textual named entity recognition, question answering, abstractive summarization, and topic classification. To support these tasks, we release three new expert-annotated Greek financial datasets and incorporate two existing resources. Our comprehensive evaluation of 22 LLMs reveals persistent challenges in Greek financial NLP, driven by linguistic complexity, domain terminology, and financial reasoning gaps. Experiment results underscore the limitations of cross-lingual transfer and the need for Greek-specific financial modeling. We publicly release Plutus-ben, Plutus-8B, and all associated datasets to promote reproducible research and advance multilingual financial NLP.

Plutus: Benchmarking Large Language Models in Low-Resource Greek Finance

Reinforcement learning with verifiable rewards (RLVR) has become a key technique for enhancing large language models (LLMs), with verification engineering playing a central role. However, best practices for RL in instruction following remain underexplored. In this work, we explore the verification challenge in RL for instruction following and propose VerIF, a verification method that combines rule-based code verification with LLM-based verification from a large reasoning model (e.g., QwQ-32B). To support this approach, we construct a high-quality instruction-following dataset, VerInstruct, containing approximately 22,000 instances with associated verification signals. We apply RL training with VerIF to two models, achieving significant improvements across several representative instruction-following benchmarks. The trained models reach state-of-the-art performance among models of comparable size and generalize well to unseen constraints. We further observe that their general capabilities remain unaffected, suggesting that RL with VerIF can be integrated into existing RL recipes to enhance overall model performance. We will release our datasets, codes, and models to facilitate future research.

VerIF: Verification Engineering for Reinforcement Learning in Instruction Following

Rapid deployment of vision-language models (VLMs) magnifies safety risks, yet most evaluations rely on artificial images. This study asks: How safe are current VLMs when confronted with meme images that ordinary users share? To investigate this question, we introduce MemeSafetyBench, a 50,430-instance benchmark pairing real meme images with both harmful and benign instructions. Using a comprehensive safety taxonomy and LLM-based instruction generation, we assess multiple VLMs across single and multi-turn interactions. We investigate how real-world memes influence harmful outputs, the mitigating effects of conversational context, and the relationship between model scale and safety metrics. Our findings demonstrate that VLMs show greater vulnerability to meme-based harmful prompts than to synthetic or typographic images. Memes significantly increase harmful responses and decrease refusals compared to text-only inputs. Though multi-turn interactions provide partial mitigation, elevated vulnerability persists. These results highlight the need for ecologically valid evaluations and stronger safety mechanisms.

Are Vision-Language Models Safe in the Wild? A Meme-Based Benchmark Study

We explore the possibility of semantic networks as a diagnostic tool for cognitive decline by using Dutch verbal fluency data to investigate the relationship between semantic networks and cognitive health. We explore construction of these networks using the contextual word embedding models BERTje (Dutch) and XLMRoBERTa (multilingual) and static embedding model FastText. In psychology, semantic networks serve as abstract representations of the semantic memory system. Semantic verbal fluency data can be used to estimate said networks. Traditionally, this is done by counting the number of raw items produced by participants in a verbal fluency task. We used the models to connect the elicited words through semantic similarity scores, and extracted three network distance metrics in addition to the traditional approach of counting. We then tested how well these metrics predict participants' cognitive health scores on the Mini-Mental State Examination (MMSE). We find that the centroid-diameter and average distance metrics are significant predictors of cognitive health scores in our data, with the traditional scoring method not reaching significance.

Downloads

Next from EMNLP 2025

Instructing Large Language Models for Low-Resource Languages: A Systematic Study for Basque

Similar lecture

AraHalluEval: A Fine-grained Hallucination Evaluation Framework for Arabic LLMs