China

With the widespread adoption of Large Language Models (LLMs) across various applications, it is empirical to ensure their fairness across all user communities. However, most LLMs are trained and evaluated on Western centric data, with little attention paid to low-resource languages and regional contexts. To address this gap, we introduce PakBBQ, a culturally and regionally adapted extension of the original Bias Benchmark for Question Answering (BBQ) dataset. PakBBQ comprises over 214 templates, 17180 QA pairs across 8 categories in both English and Urdu, covering eight bias dimensions including age, disability, appearance, gender, socio-economic status, religious, regional affiliation, and language formality that are relevant in Pakistan. We evaluate multiple multilingual LLMs under both ambiguous and explicitly disambiguated contexts, as well as negative versus non negative question framings. Our experiments reveal (i) an average accuracy gain of 12% with disambiguation, (ii) consistently stronger counter bias behaviors in Urdu than in English, and (iii) marked framing effects that reduce stereotypical responses when questions are posed negatively. These findings highlight the importance of contextualized benchmarks and simple prompt engineering strategies for bias mitigation in low resource settings.

EMNLP 2025

PakBBQ: A Culturally Adapted Bias Benchmark for QA

poster

## 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*

[**Link to Conference Handbook**](https://drive.google.com/file/d/1johU5QqVVYO4RfH7QcIORr7qrVBdzdwC/view?usp=sharing)





<br>

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

Recent research has shown that carefully crafted jailbreak inputs can induce large language models to produce harmful outputs, despite safety measures such as alignment. It is important to anticipate the range of potential Jailbreak attacks to guide effective defenses and accurate assessment of model safety. In this paper, we present a new approach for generating highly effective Jailbreak attacks that manipulate the attention of the model to selectively strengthen or weaken attention among different parts of the prompt. By harnessing attention loss, we develop more effective jailbreak attacks, that are also transferrable. The attacks amplify the success rate of existing Jailbreak algorithms, including GCG, AutoDAN, and ReNeLLM, while lowering their generation cost (for example, the amplified GCG attack achieves 91.2% ASR, vs. 67.9% for the original attack on Llama2-7B-chat/AdvBench, using less than a third of the generation time).

Attention Eclipse: Manipulating Attention to Bypass LLM Safety-Alignment

As one of the world's most populous countries, with 700 languages spoken, Indonesia is behind in terms of NLP progress. We introduce LORAXBENCH, a benchmark that focuses on low-resource languages of Indonesia and covers 6 diverse tasks: reading comprehension, open-domain QA, language inference, causal reasoning, translation, and cultural QA. Our dataset cover 20 languages, with the addition of two formality registers for three languages. We evaluate a diverse set of multilingual and region-focused LLMs and found that this benchmark is challenging. We note a visible discrepancy between performance in Indonesian and other languages, especially the low-resource ones. There is no clear lead when using a region-specific model as opposed to the general multilingual model. Lastly, we show that a change in register affects model performance, especially with registers not commonly found in social media, such as high-level politeness `Krama' Javanese.

LORAXBENCH: A Multitask, Multilingual Benchmark Suite for 20 Indonesian Languages

Compact dual-encoder models are widely used for retrieval owing to their efficiency and scalability. However, such models often underperform compared to their Large Language Model (LLM)-based retrieval counterparts, likely due to their limited world knowledge. While LLM-based data augmentation has been proposed as a strategy to bridge this performance gap, there is insufficient understanding of its effectiveness and scalability to real-world retrieval problems. Existing research does not systematically explore key factors such as the optimal augmentation scale, the necessity of using large augmentation models, and whether diverse augmentations improve generalization, particularly in out-of-distribution (OOD) settings. This work presents a comprehensive study of the effectiveness of LLM augmentation for retrieval, comprising over 100 distinct experimental settings of retrieval models, augmentation models and augmentation strategies. We find that, while augmentation enhances retrieval performance, its benefits diminish beyond a certain scale, even with diverse augmentation strategies. Surprisingly, we observe that augmentation with smaller LLMs can achieve performance competitive with larger augmentation models. Moreover, we examine how augmentation effectiveness varies with retrieval model pre-training, revealing that augmentation provides the most benefit to models which are not well pre-trained. Our insights pave the way for more judicious and efficient augmentation strategies, thus enabling informed decisions and maximizing retrieval performance while being more cost-effective.

Evaluating the Effectiveness and Scalability of LLM-Based Data Augmentation for Retrieval

Recent advances in large language models (LLMs) have demonstrated significant promise in document understanding and question-answering. Despite the progress, existing approaches can only process short documents due to limited context length or fail to fully leverage multi-modal information. In this work, we introduce DocAgent, a multi-agent framework for long-context document understanding that imitates human reading practice. Specifically, we first extract a structured, tree-formatted outline from documents to help agents identify relevant sections efficiently. Further, we develop an interactive reading interface that enables agents to query and retrieve various types of content dynamically. To ensure answer reliability, we introduce a reviewer agent that cross-checks responses using complementary sources and maintains a task-agnostic memory bank to facilitate knowledge sharing across tasks. We evaluate our method on two long-context document understanding benchmarks, where it bridges the gap to human-level performance by surpassing competitive baselines, while maintaining a short context length.

DocAgent: An Agentic Framework for Multi-Modal Long-Context Document Understanding

Software debugging is a time-consuming endeavor involving a series of steps, such as fault localization and patch generation, each requiring thorough analysis and a deep understanding of the underlying logic. While large language models (LLMs) demonstrate promising potential in coding tasks, their performance in debugging remains limited. Current LLM-based methods often focus on isolated steps and struggle with complex bugs. In this paper, we propose the first end-to-end framework, UniDebugger, for unified debugging through multi-agent synergy. It mimics the entire cognitive processes of developers, with each agent specialized as a particular component of this process rather than mirroring the actions of an independent expert as in previous multi-agent systems. Agents are coordinated through a three-level design, following a cognitive model of debugging, allowing adaptive handling of bugs with varying complexities. Experiments on extensive benchmarks demonstrate that UniDebugger significantly outperforms state-of-the-art repair methods, fixing 1.25x to 2.56x bugs on the repo-level benchmark, Defects4J. This performance is achieved without requiring ground-truth root-cause code statements, unlike the baselines. Our source code is available on an anonymous link: https://github.com/AcceptePapier/UniDebugger.

UniDebugger: Hierarchical Multi-Agent Framework for Unified Software Debugging

An Agentic AI Workflow (AAW), also known as an LLM-based multi-agent system, is an autonomous system that assembles several LLM-based agents to work collaboratively towards a shared goal. The high autonomy, widespread adoption, and growing interest in such AAWs highlight the need for a deeper understanding of their operations, from both quality and security aspects. To this day, there are no existing methods to assess the influence of each agent on the AAW's final output. Adopting techniques from related fields is not feasible since existing methods perform only static structural analysis, which is unsuitable for inference time execution. We present Counterfactual-based Agent Influence Ranker (CAIR) - the first method for assessing the influence level of each agent on the AAW's output and determining which agents are the most influential. By performing counterfactual analysis, CAIR provides a task-agnostic analysis that can be used both offline and at inference time. We evaluate CAIR using an AAWs dataset of our creation, containing 30 different use cases with 230 different functionalities. Our evaluation showed that CAIR produces consistent rankings, outperforms baseline methods, and can easily enhance the effectiveness and relevancy of downstream tasks.

CAIR: Counterfactual-based Agent Influence Ranker for Agentic AI Workflows

The effectiveness of Large Language Models (LLMs) in solving tasks depends significantly on the quality of their instructions, which often require substantial human effort to craft. This underscores the need for automated instruction optimization. However, optimizing instructions is particularly challenging when working with black-box LLMs, where model parameters and gradients are inaccessible. We introduce ACING, an actor-critic reinforcement learning framework that formulates instruction optimization as a stateless, continuous-action problem, enabling exploration of infinite instruction spaces using only black-box feedback. ACING automatically discovers prompts that outperform human-written prompts in 76\% of instruction-induction tasks, with gains of up to 33 points and a 10-point median improvement over the best automatic baseline in 33 tasks spanning instruction-induction, summarization, and chain-of-thought reasoning. Extensive ablations highlight its robustness and efficiency. Code will be released as open source.

ACING: Actor-Critic for Instruction Learning in Black-Box LLMs

Automated Essay Scoring (AES) systems now attain near–human agreement on public benchmarks, yet real-world adoption—especially in high-stakes examinations—remains limited. A principal obstacle is that most models output a single score without any accompanying measure of confidence or explanation. We address this gap with conformal prediction, a distribution-free wrapper that equips any classifier with set-valued outputs enjoying formal coverage guarantees. Two open-weight large language models—Llama-3 8B and Qwen-2.5 3B—are fine-tuned on three diverse corpora (ASAP, TOEFL11, Cambridge-FCE) and calibrated at a 90% risk level. Reliability is assessed with UAcc, an uncertainty-aware accuracy that rewards models for being both correct and concise. To our knowledge, this is the first work to combine conformal prediction and UAcc for essay scoring. The calibrated models consistently meet the coverage target while keeping prediction sets compact, demonstrating that trustworthy, uncertainty-aware AES is already feasible with mid-sized, open source LLMs and paving the way for safer human-in the-loop marking.

Beyond the Score: Uncertainty-Calibrated LLMs for Automated Essay Assessment

As large language models continue to scale, computational costs and resource consumption have emerged as significant challenges. While existing sparsification methods like pruning reduce computational overhead, they risk losing model knowledge through parameter removal. This paper proposes DSMoE (Dynamic Sparse Mixture-of-Experts), a novel approach that achieves sparsification by partitioning pre-trained FFN layers into computational blocks. We implement adaptive expert routing using sigmoid activation and straight-through estimators, enabling tokens to flexibly access different aspects of model knowledge based on input complexity. Additionally, we introduce a sparsity loss term to balance performance and computational efficiency. Extensive experiments on LLaMA models demonstrate that under equivalent computational constraints, DSMoE achieves superior performance compared to existing pruning and MoE approaches across language modeling and downstream tasks, particularly excelling in generation tasks. Analysis reveals that DSMoE learns distinctive layerwise activation patterns, providing new insights for future MoE architecture design.

DSMoE: Matrix-Partitioned Experts with Dynamic Routing for Computation-Efficient Dense LLMs

Large multimodal models (LMMs) have recently gained attention due to their effectiveness to understand and generate descriptions of visual content. Most existing LMMs are in English language. While few recent works explore multilingual image LMMs, to the best of our knowledge, moving beyond the English language for cultural and linguistic inclusivity is yet to be investigated in the context of video LMMs. In pursuit of more inclusive video LMMs, we introduce a multilingual Video LMM benchmark, named ViMUL-Bench, to evaluate Video LMMs across 14 languages, including both low- and high-resource languages: English, Chinese, Spanish, French, German, Hindi, Arabic, Russian, Bengali, Urdu, Sinhala, Tamil, Swedish, and Japanese. Our ViMUL-Bench is designed to rigorously test video LMMs across 15 categories including eight culturally diverse categories, ranging from lifestyles and festivals to foods and rituals and from local landmarks to prominent cultural personalities. ViMUL-Bench comprises both open-ended (short and long-form) and multiple-choice questions spanning various video durations (short, medium, and long) with 8k samples that are manually verified by native language speakers. In addition, we also introduce a machine translated multilingual video training set comprising 1.2 million samples and develop a simple multilingual video LMM, named ViMUL, that is shown to provide a better tradeoff between high-and low-resource languages for video understanding. We hope our ViMUL-Bench and multilingual video LMM along with a large-scale multilingual video training set will help ease future research in developing cultural and linguistic inclusive multilingual video LMMs. Our proposed benchmark, video LMM and training data will be publicly released.

Downloads

Next from EMNLP 2025

Attention Eclipse: Manipulating Attention to Bypass LLM Safety-Alignment

Stay up to date with the latest Underline news!

PRESENTATIONS

CONFERENCES

COMPANY

RESOURCES