Morocco

Reliable uncertainty quantification (UQ) is essential when employing large language models (LLMs) in high-risk domains such as clinical question answering (QA). In this work, we evaluate uncertainty estimation methods for clinical QA focusing, for the first time, on eleven clinical specialties and six question types, and across ten open-source LLMs (general-purpose, biomedical, and reasoning models). We analyze score-based UQ methods, present a case study introducing a novel lightweight method based on behavioral features derived from reasoning-oriented models, and examine conformal prediction as a complementary set-based approach. Our findings reveal that uncertainty reliability is not a monolithic property, but one that depends on clinical specialty and question type due to shifts in calibration and discrimination. Our results highlight the need to select or ensemble models based on their distinct, complementary strengths and clinical use.

EACL 2026

Mind the Gap: Benchmarking LLM Uncertainty and Calibration with Specialty-Aware Clinical QA and Reasoning-Based Behavioural Features

calibration and discrimination

llms

uncertainty quantification

reasoning

clinical nlp

It is my great pleasure and honor to welcome you to the 64th Annual Meeting of the Association for Computational Linguistics (ACL 2026), held in San Diego, California, USA, July 2–7, 2026. ACL 2026 continues our fields 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.

It is exciting to see the many developments in the field reflected at our conference, such as the advancement of large language models (born out of our community) into many application directions. This is not without challenges, partly due to unsolved problems, and partly due to misuse of the technology. All this is reflected well at ACL 2026. The conference also reflects that the field is not only about advancing technology, but also the science of fundamentally understanding language and how it is use by many different speaker communities. <br>
[Full Message from General Chair Philipp Koehn](https://drive.google.com/file/d/1QiJFOmMjvDgsDo3i1XzlQnImxe3RU_Cy/view?usp=sharing) <br><br><br>
<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://drive.google.com/file/d/1QiJFOmMjvDgsDo3i1XzlQnImxe3RU_Cy/view?usp=sharing')">Conference Handbook</button></html>
<br><br><br><br>

To access this event page, you need to log in with the **email address you registered with**. <br>Access credentials will be sent to your email from Underline -  subject line "Welcome to ACL 2026". Please be sure to check your spam email folder if you do not see an email confirmation right away.

To access this event page, you are required to register.
Please complete your registration to continue.

We recommend reading [**the registration information**](https://2026.aclweb.org/registration/) first.

Registration Required

ACL 2026

The 64th Annual Meeting of the Association for Computational Linguistics (ACL 2026) will take place in San Diego, California, from July 2 - 7, 2026.

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)*<br><br>

<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>
<br><br><br>

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

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. 

Retrieval-Augmented Generation (RAG) systems depend critically on retrieval quality to enable accurate, contextually relevant LLM responses. While LLMs excel at synthesis, their RAG performance is bottlenecked by document relevance. We evaluate advanced retrieval techniques including embedding model comparison, Reciprocal Rank Fusion (RRF), embedding concatenation and list-wise and adaptive LLM-based re-ranking, demonstrating that zero-shot LLMs outperform traditional cross-encoders in identifying high-relevance passages. We also explore context-aware embeddings, diverse chunking strategies, and model fine-tuning. All methods are rigorously evaluated on a proprietary dataset powering our deployed production chatbot, with validation on three public benchmarks: FiQA, HotpotQA, and SciDocs. Results show consistent gains in Recall@10, closing the gap with Recall@50 and yielding actionable pipeline recommendations. By prioritizing retrieval enhancements, we significantly elevate downstream LLM response quality in real-world, customer-facing applications.

Retrieval Enhancements for RAG: Insights from a Deployed Customer Support Chatbot

Generative Retrieval (GR) maps queries to documents by generating discrete identifiers (DocIDs). However, offline DocID assignment and constrained decoding often prevent GR from capturing query-specific intent, especially when documents express multiple or unseen intents (i.e., intent misalignment). We introduce Dynamic Docid Decoding (D3), an inference-time mechanism that adaptively refines DocIDs through delayed, query-informed identifier expansion. D3 uses (a) verification to detect intent misalignment and (b) dynamic decoding to extend DocIDs with query-aligned tokens, even those absent from the pre-indexed vocabulary, enabling plug-and-play DocID expansion beyond the static vocabulary while adding minimal overhead. Experiments on NQ320k and MS-MARCO show that D3 consistently improves retrieval accuracy, especially on unseen and multi-intent documents, across various GR models, including a +2.4\%p nDCG@10 gain on the state-of-the-art model.

D3: Dynamic Docid Decoding for Multi-Intent Generative Retrieval

Arabic calligraphy represents one of the richest visual traditions of the Arabic language, blending linguistic meaning with artistic form. Although multimodal models have advanced across languages, their ability to process Arabic script, especially in artistic and stylized calligraphic forms, remains largely unexplored. To address this gap, we present DuwatBench, a benchmark of 1,050 curated samples containing about 1,400 words across 6 classical and modern calligraphic styles, each paired with sentence-level detection annotations. The dataset reflects real-world challenges in Arabic writing, such as complex stroke patterns, dense ligatures, and stylistic variations that often challenge standard text recognition systems. Using DuwatBench, we evaluated 13 leading Arabic and multilingual multimodal models and show that while they perform well in clean text, they struggle with calligraphic variation, artistic distortions, and precise visualвЂ“text alignment. By publicly releasing DuwatBench and its annotations, we aim to advance culturally grounded multimodal research, foster fair inclusion of Arabic language and visual heritage in AI systems, and support continued progress in this area. Our benchmark and the evaluation suit will be publicly available.

DuwatBench: Bridging Language and Visual Heritage through an Arabic Calligraphy Benchmark for Multimodal Understanding

For socially sensitive tasks like hate speech detection, the quality of explanations from Large Language Models (LLMs) is crucial for factors like user trust and model alignment. While Persona prompting (PP) is increasingly used as a way to steer model towards user-specific generation, its effect on model rationales remains underexplored. We investigate how LLM-generated rationales vary when conditioned on different simulated demographic personas. Using datasets annotated with word-level rationales, we measure agreement with human annotations from different demographic groups, and assess the impact of PP on model bias and human alignment. Our evaluation across three LLMs results reveals three key findings: (1) PP improving classification on the most subjective task (hate speech) but degrading rationale quality. (2) Simulated personas fail to align with their real-world demographic counterparts, and high inter-persona agreement shows models are resistant to significant steering. (3) Models exhibit consistent demographic biases and a strong tendency to over-flag content as harmful, regardless of PP. Our findings reveal a critical trade-off: while PP can improve classification in socially-sensitive tasks, it often comes at the cost of rationale quality and fails to mitigate underlying biases, urging caution in its application.

Persona Prompting as a Lens on LLM Social Reasoning

Medical reasoning in large language models seeks to replicate clinicians' cognitive processes in interpreting patient data and making diagnostic decisions. However, widely used benchmarksвЂ”such as MedQA, MedMCQA, and PubMedQAвЂ”mix questions that require multi-step reasoning with those answerable through factual recall, complicating evaluation. We demonstrate this by training a PubMedBERT-based classifier on expert-curated labels and applying it to 11 widely used biomedical QA benchmarks, where we find that only 32.8% of the questions require multi-step reasoning, indicating that current evaluations largely measure recall. This stratified evaluation of biomedical models (HuatuoGPT-o1, MedReason, m1) and general-domain models (DeepSeek-R1, o4-mini, Qwen3) reveals consistently lower performance on reasoning versus knowledge (e.g., HuatuoGPT-o1: 56.9% vs. 44.8%). Beyond accuracy, we assess robustness through adversarial evaluations in which models are prefixed with uncertainty-inducing statements; biomedical reasoning models degrade sharply in this setting (e.g., MedReason: 50.4% в†’ 24.4%), with declines especially pronounced on reasoning-heavy questions. Finally, we show that fine-tuning on high-quality reasoning examples augmented with adversarial traces, followed by reinforcement learning with GRPO, improves both robustness and accuracy across knowledge and reasoning subsets.

Reasoning or Knowledge: Stratified Evaluation of Biomedical LLMs

Sparse Autoencoders (SAEs) are powerful tools for interpreting neural representations, yet their use in audio remains underexplored. We train SAEs across all encoder layers of Whisper and HuBERT, provide an extensive evaluation of their stability, interpretability, and show their practical utility. Over 50% of the features remain consistent across random seeds, and reconstruction quality is preserved. SAE features capture general acoustic and semantic information as well as specific events, including environmental noises and paralinguistic sounds (e.g. laughter, whispering) and disentangle them effectively, requiring removal of only 19-27% of features to erase a concept. Feature steering reduces Whisper’s false speech detections by 70% with negligible WER increase, demonstrating real-world applicability. Finally, we find SAE features correlated with human EEG activity during speech perception, indicating alignment with human neural processing. The code and checkpoints are available at https://github.com/audiosae/audiosae_demo.

AudioSAE: Towards Understanding of Audio-Processing Models with Sparse AutoEncoders

Despite their impressive performance, vision-language models (VLMs) still struggle on culturally situated inputs. To understand how VLMs process culturally grounded information, we study the presence of culture-sensitive neurons, i.e. neurons whose activations show preferential sensitivity to inputs associated with particular cultural contexts. We examine whether such neurons are important for culturally diverse visual question answering and where they are located. Using the CVQA benchmark, we identify neurons of culture selectivity and perform causal tests by deactivating the neurons flagged by different identification methods. Experiments on three VLMs across 25 cultural groups demonstrate the existence of neurons whose ablation disproportionately harms performance on questions about the corresponding cultures, while having minimal effects on others. Moreover, we propose a new margin-based selector: Contrastive Activation Selection (CAS), and show that it outperforms existing probability- and entropy-based methods in identifying culture-sensitive neurons. Finally, our layer-wise analyses reveals that such neurons tend to cluster in certain decoder layers. Overall, our findings shed new light on the internal organization of multimodal representations.

Finding Culture-Sensitive Neurons in Vision-Language Models

Multiple-Choice Questions (MCQs) are often used to assess knowledge, reasoning abilities, and even values encoded in large language models (LLMs). While the effect multilingualism has been studied on LLM factual recall, few studies have taken interest in language-induced variation in value-laden MCQ responses. This paper seeks to bridge that gap: do multilingual LLMs behave like polyglots, i.e. answer the same question the same way across languages, or do they answer value-laden MCQs depending on the language, like a multitude of monolingual models? Unlike prior work relying on machine translation or ad hoc prompts, we release a new corpus solely comprising human-translated survey questions aligned in 8 European languages. We administer a subset of those questions to about forty multilingual LLMs of various sizes, manufacturers and alignment-fine-tuning status under comprehensive, controlled prompt variations including answer order, symbol type, and tail character. Our results show that while instruction-tuned, larger models display higher consistency overall, that consistency varies greatly across questions, with certain MCQs eliciting total agreement within and across models while other leave LLM answers evenly split. Language-specific behavior seems to arise in all of the most consistent models, but only on certain questions, warranting a further study of the selective effect of preference fine-tuning.

Polyglots or Multitudes? Multilingual LLM Answers to Value-laden Multiple-Choice Questions

Persona-assigned large language models (LLMs) are used in domains such as education, healthcare, and sociodemographic simulation. Yet, they are typically evaluated only in short, single-round settings that do not reflect real-world usage. We introduce an evaluation protocol that combines long persona dialogues (over 100 rounds) and evaluation datasets to create dialogue-conditioned benchmarks that can robustly measure long-context effects. We then investigate the effects of dialogue length on persona fidelity, instruction-following, and safety of seven state-of-the-art open- and closed-weight LLMs. We find that persona fidelity degrades over the course of dialogues, especially in goal-oriented conversations, where models must sustain both persona fidelity and instruction following. We identify a trade-off between fidelity and instruction following, with non-persona baselines initially outperforming persona-assigned models; as dialogues progress and fidelity fades, persona responses become increasingly similar to baseline responses. Our findings highlight the fragility of persona applications in extended interactions and our work provides a protocol to systematically measure such failures.

Persistent Personas? Role-Playing, Instruction Following, and Safety in Extended Interactions

The potential misuse and misalignment of language models (LMs) is a central safety concern. This work presents Self-Destruct, a novel mechanism to restrict specific behaviors in LMs by leveraging overlooked properties of the underlying hardware. We observe that the LM frameworks use limited-precision formats (e.g., FP32), which are vulnerable to overflow errors during matrix multiplications. Exploiting this property, Self-Destruct replaces selected weights in pre-trained LM layers with values that act as traps, triggering a system error only when the model engages in targeted behaviors, such as harmful text generation, while leaving normal functionality unaffected. Unlike post-hoc filters, this safeguard is embedded directly within the model, introduces neither inference overhead nor auxiliary models, and requires only a set of examples for calibration. Experiments on Llama-3.2 and Qwen-2.5 demonstrate that Self-Destruct provides competitive protection against jailbreak attacks while preserving accuracy on standard benchmarks. Our results highlight the potential of hardware-aware safeguards as an efficient, low-overhead complement to existing LM defenses.

Downloads

Next from EACL 2026

Retrieval Enhancements for RAG: Insights from a Deployed Customer Support Chatbot

Similar lecture

Uncertainty Quantification of Large Language Models through Multiple Uncertainty Sources

.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 EACL 2026

Retrieval Enhancements for RAG: Insights from a Deployed Customer Support Chatbot

Similar lecture

Uncertainty Quantification of Large Language Models through Multiple Uncertainty Sources

Downloads