Morocco

Does LLM &quot;reasoning&quot; rely on logic or just surface recall? While memorization inflates performance in math, real-world tasks like story understanding require selectively constrained memorization.

We introduce a two-tier framework—Inductive (soft-steering) and Restrictive (hard-constraint)—to evaluate LLMs (GPT-4o, LLaMA3.3, DeepSeek V3) across six character-centric benchmarks. Our findings reveal a critical gap:

The &quot;Recall Trap&quot;: Accuracy drops by up to 45.2% when verbatim memory access is restricted.

The Path Forward: Inductive prompting effectively aligns models toward &quot;gist&quot; reasoning rather than rote memorization.

This work provides a new lens for distinguishing true semantic abstraction from simple data retrieval.

EACL 2026

Beyond Math: Stories as a Testbed for Memorization-Constrained Reasoning in LLMs

memorization evaluation

llm memorization

character understanding

## 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>

EMNLP 2025

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

Does LLM "reasoning" rely on logic or just surface recall? While memorization inflates performance in math, real-world tasks like story understanding require selectively constrained memorization.

We introduce a two-tier framework—Inductive (soft-steering) and Restrictive (hard-constraint)—to evaluate LLMs (GPT-4o, LLaMA3.3, DeepSeek V3) across six character-centric benchmarks. Our findings reveal a critical gap:

The "Recall Trap": Accuracy drops by up to 45.2% when verbatim memory access is restricted.

The Path Forward: Inductive prompting effectively aligns models toward "gist" reasoning rather than rote memorization.

This work provides a new lens for distinguishing true semantic abstraction from simple data retrieval.

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

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. 

Membership inference attacks (MIAs) reveal whether specific data was used to train machine learning models, serving as important tools for privacy auditing and compliance assessment. Recent studies have reported that MIAs perform only marginally better than random guessing against large language models, suggesting that modern pre-training approaches with massive datasets may be free from privacy leakage risks. Our work offers a complementary perspective to these findings by exploring how examining LLMs' internal representations, rather than just their outputs, may provide additional insights into potential membership inference signals. Our framework, \emph{memTrace}, follows what we call \enquote{neural breadcrumbs} extracting informative signals from transformer hidden states and attention patterns as they process candidate sequences. By analyzing layer-wise representation dynamics, attention distribution characteristics, and cross-layer transition patterns, we detect potential memorization fingerprints that traditional loss-based approaches may not capture. This approach yields strong membership detection across several model families achieving average AUC scores of 0.85 on popular MIA benchmarks. Our findings suggest that internal model behaviors can reveal aspects of training data exposure even when output-based signals appear protected, highlighting the need for further research into membership privacy and the development of more robust privacy-preserving training techniques for large language models.

Neural Breadcrumbs: Membership Inference Attacks on LLMs Through Hidden State and Attention Pattern Analysis

Recent advances in multimodal large language models (MLLMs) offer a promising approach for natural language-based scene change queries in virtual reality (VR). Prior work on applying MLLMs for object state understanding has focused on egocentric videos that capture the camera wearer's interactions with objects. However, object state changes may occur in the background without direct user interaction, lacking explicit motion cues and making them difficult to detect. Moreover, no benchmark exists for evaluating this challenging scenario. To address these challenges, we introduce ObjChangeVR-Dataset, specifically for benchmarking the question-answering task of object state change. We also propose ObjChangeVR, a framework that combines viewpoint-aware and temporal-based retrieval to identify relevant frames, along with cross-view reasoning that reconciles inconsistent evidence from multiple viewpoints. Extensive experiments demonstrate that ObjChangeVR significantly outperforms baseline approaches across multiple state-of-the-art MLLMs.

ObjChangeVR: Object State Change Reasoning from Continuous Egocentric Views in VR Environments

Large Language Models (LLMs) fail at reasoning when their parametric knowledge is outdated or incorrect. A common solution for mitigating outdated or incorrect information in models is to provide updated facts in-context or by knowledge editing, but such updates can conflict with LLMs' parametric knowledge and limit their reasoning. Current benchmarks for this problem, however, largely focus only on single knowledge updates and fact recall in isolated scenarios. In this work, we introduce TКЂбґЂбґ„бґ‹ (*Testing Reasoning Amid Conflicting Knowledge*), a new benchmark for studying how LLMs propagate new knowledge through multi-step reasoning when it conflicts with the model's initial parametric knowledge. Spanning three reasoning-intensive scenarios (WIKI, CODE, and MATH), TКЂбґЂбґ„бґ‹ introduces multiple, realistic conflicts to mirror real-world complexity. Our results on TКЂбґЂбґ„бґ‹ reveal that providing updated facts yields limited performance gains and even worsens performance, compared with no facts provided. Performance further degrades with more updated facts. This failure stems from both inability to faithfully integrate updated facts and flawed reasoning even when knowledge is integrated. TКЂбґЂбґ„бґ‹ provides a rigorous new benchmark to measure and guide future progress on propagating conflicting knowledge in multi-step reasoning.

Tracking the Limits of Knowledge Propagation: How LLMs Fail at Multi-Step Reasoning with Conflicting Knowledge

Understanding emotion from speech requires sensitivity to both lexical and acoustic cues. However, it remains unclear whether large audio language models (LALMs) genuinely process acoustic information or rely primarily on lexical contents. We present LISTEN (Lexical vs. Acoustic Speech Test for Emotion in Narratives), a controlled benchmark designed to disentangle lexical reliance from acoustic sensitivity in emotion understanding. Across evaluations of six state-of-the-art LALMs, we observe a consistent lexical dominance. Models predict вЂњneutralвЂќ when lexical cues are neutral or absent, show limited gains under cue alignment, and fail to classify distinct emotions under cue conflict. In paralinguistic settings, performance approaches chance. These results indicate that current LALMs largely вЂњtranscribeвЂќ rather than вЂњlisten,вЂќ relying heavily on lexical semantics while underutilizing acoustic cues. LISTEN offers a principled framework for assessing emotion understanding in multimodal models.

Do Audio LLMs Really LISTEN, or Just Transcribe? Measuring Lexical vs. Acoustic Emotion Cues Reliance

Recent advances in self-supervised learning (SSL) have led to powerful speech representation models, yet their robustness in real-world conversational settings remains largely untested. Most existing benchmarks focus on clean, single-speaker, and single-channel audio, failing to reflect the complexities of natural human interactionвЂ”where overlapping speech, background noise, and reverberation are the norm. To bridge these critical gaps, we present the Conversational Speech Processing Benchmark (CSPB), a new benchmark designed to assess the robustness of SSL speech models in realistic conversational scenarios. CSPB is constructed from four multi-party datasetsвЂ”AMI, AliMeeting, MMCSG, and DiPCoвЂ”and supports both single-channel and multi-channel evaluation. By releasing CSPB as an open-source toolkit, we aim to establish a unified framework for evaluating and advancing robust, spatially-aware self-supervised speech models.

CSPB: Conversational Speech Processing Benchmark for Self-supervised Speech Models

Text anonymization is a critical task for enabling research and development in high-stakes domains containing private data, like medicine, law, and social services. While much research has focused on redacting sensitive content from text, substantially less work has focused on what to replace redacted content with, which can enhance privacy and becomes increasingly important with greater levels of redaction. In this work, we formulate predicting replacements for sensitive spans as a research task with principled use-inspired evaluation criteria. We further propose a multi-token completion method for accomplishing this task that is designed to preserve consistency with low compute requirements, thus facilitating practitioners to anonymize data locally before sharing it externally. Human and automated annotations demonstrate that our approach produces more realistic text and better preserves utility than alternative infilling methods and differentially private mechanisms across multiple domains without retraining. Overall, our work explores the under-studied task of what to replace redacted content with and contributes grounded evaluations capturing utility, facilitating future work.

Multi-Token Completion for Text Anonymization

Large audio-language models increasingly operate on raw speech inputs, enabling more seamless integration across domains such as voice assistants, education, and clinical triage. This transition, however, introduces a distinct class of vulnerabilities that remain largely uncharacterized. We examine the security implications of this modality shift by designing a text-to-audio jailbreak that embeds disallowed directives within a narrative-style audio stream. The attack leverages an advanced instruction-following text-to-speech (TTS) model to exploit structural and acoustic properties, thereby circumventing safety mechanisms primarily calibrated for text. When delivered through synthetic speech, the narrative format elicits restricted outputs from state-of-the-art models, including Gemini 2.0 Flash, achieving a 98.26% success rate that substantially exceeds text-only baselines. These results highlight the need for safety frameworks that jointly reason over linguistic and paralinguistic representations, particularly as speech-based interfaces become more prevalent.

Now You Hear Me: Audio Narrative Attacks Against Large AudioвЂ“Language Models

Sparse autoencoders (SAEs) are commonly used to interpret the internal activations of large language models (LLMs) by mapping them to human-interpretable concept representations. While existing evaluations of SAEs focus on metrics such as the reconstruction-sparsity tradeoff, human (auto-)interpretability, and feature disentanglement, they overlook a critical aspect: the robustness of concept representations to input perturbations. We argue that robustness must be a fundamental consideration for concept representations, reflecting the fidelity of concept labeling. To this end, we formulate robustness quantification as input-space optimization problems and develop a comprehensive evaluation framework featuring realistic scenarios in which adversarial perturbations are crafted to manipulate SAE representations. Empirically, we find that tiny adversarial input perturbations can effectively manipulate concept-based interpretations in most scenarios without notably affecting the base LLM's activations. Overall, our results suggest that SAE concept representations are fragile and without further denoising or postprocessing they might be ill-suited for applications in model monitoring and oversight.

Evaluating Adversarial Robustness of Concept Representations in Sparse Autoencoders

Evaluating the quality of children's utterances in adult-child dialogue remains challenging due to insufficient context-sensitive metrics. Common proxies such as Mean Length of Utterance (MLU), lexical diversity (vocd-D), and readability indices (Flesch-Kincaid Grade Level, Gunning Fog Index) are dominated by length and ignore conversational context, missing aspects of response quality such as reasoning depth, topic maintenance, and discourse planning. We introduce an LLM-as-a-judge framework that first classifies the Previous Adult Utterance Type and then scores the child's response along two axes: Expansion (contextual elaboration and inferential depth) and Independence (the child's contribution to advancing the discourse). These axes reflect fundamental dimensions in child language development, where Expansion captures elaboration, clause combining, and causal and contrastive connectives. Independence captures initiative, topic control, decreasing reliance on adult scaffolding through growing self-regulation, and audience design. We establish developmental validity by showing age-related patterns and demonstrate predictive value by improving age estimation over common baselines. We further confirm semantic sensitivity by detecting differences tied to discourse relations. Our metrics align with human judgments, enabling large-scale evaluation. This shifts child utterance assessment from simply measuring length to evaluating how meaningfully the child's speech contributes to and advances the conversation within its context.

Beyond Length: Context-Aware Expansion and Independence as Developmentally Sensitive Evaluation in Child Utterances

Current evaluation of large language models (LLMs) overwhelmingly prioritizes accuracy; however, in real-world and safety-critical applications, the ability to abstain when uncertain is equally vital for trustworthy deployment. We introduce a unified benchmark and evaluation protocol for abstention in medical multiple-choice question answering (MCQA), integrating conformal prediction, adversarial question perturbations, and explicit abstention options. Our systematic evaluation of both open- and closed-source LLMs reveals that even state-of-the-art, high-accuracy models often fail to abstain when uncertain. Notably, providing explicit abstention options consistently increases model uncertainty and safer abstention, far more than input perturbations, while scaling model size or advanced prompting brings little improvement. These findings highlight the central role of abstention mechanisms for trustworthy LLM deployment and offer practical guidance for improving safety in high-stakes applications.

Downloads

Next from EACL 2026

Neural Breadcrumbs: Membership Inference Attacks on LLMs Through Hidden State and Attention Pattern Analysis

Similar lecture

Shared Path: Unraveling Memorization in Multilingual LLMs through Language Similarities

.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

Neural Breadcrumbs: Membership Inference Attacks on LLMs Through Hidden State and Attention Pattern Analysis

Similar lecture

Shared Path: Unraveling Memorization in Multilingual LLMs through Language Similarities

Downloads