Morocco

In-context learning is governed by both temporal and semantic relationships, shaping how Large Language Models (LLMs) retrieve contextual information. Analogous to human episodic memory, where the retrieval of specific events is enabled by separating events that happened at different times, this work probes the ability of various pretrained LLMs, including transformer and state-space models, to differentiate and retrieve temporally separated events. Specifically, we prompted models with sequences containing multiple presentations of the same token, which reappears at the sequence end. By fixing the positions of these repeated tokens and permuting all others, we removed semantic confounds and isolated temporal effects on next-token prediction. Across diverse sequences, models consistently placed the highest probabilities on tokens following a repeated token, but with a notable bias for those nearest the beginning or end of the input. An ablation experiment linked this phenomenon in transformers to induction heads. Extending the analysis to unique semantic contexts with partial overlap further demonstrated that memories embedded in the middle of a prompt are retrieved less reliably. Despite architectural differences, state-space and transformer models showed comparable temporal biases. Our findings deepen the understanding of temporal biases in in-context learning and offer an illustration of how these biases can enable temporal separation and episodic retrieval.

EACL 2026

Beyond Semantics: How Temporal Biases Shapes Retrieval in Transformer and State-Space Models

episodic-memory

state-space-models

transformers

### Welcome to 2025 Annual Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics

Welcome to the 2025 meeting of the Nations of the Americas Chapter of the Association for Computational Linguistics! I am proud to help organize the first NAACL conference to carry the new name of our organization, one that emphasizes inclusion for all of the Americas. I am also pleased to welcome you to Albuquerque, New Mexico, a state whose unique blend of cultural influences will make for an excellent backdrop for NAACL 2025, especially with this year’s special theme on NLP in a Multicultural World. 
**[Continue reading...](https://drive.google.com/file/d/1jX-qGhqVSZZCIrAnJaz798pflu5Irrdn/view?usp=sharing)**

*- Colin Cherry, Google, NAACL 2025 General Chair* 

[![](https://assets.underline.io/markdown_image/1/image/b087f8a4dc5816d6e1a6514e59c59ac3.png)](https://drive.google.com/file/d/1T96GzPqObXrMln2BMByCSXTSizjTg69P/view?usp=sharing)

You need to log in with the email address you registered with. Access credentials have been sent to your email. 

Please be sure to check your spam and other email folders if you do not see an email confirmation right away.

Please log in to explore this event.

To access NAACL 2025 event page you are required to register. Please follow [**this link**](https://2025.naacl.org/registration/registration/) to register. Access will depend on your registration type.

Please register!

NAACL 2025

Welcome to 2025 Annual Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics

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. 

The advent of complex, interconnected long-horizon LLM systems has made it incredibly tricky to identify where and when these systems break down. Evaluation capabilities that currently exist today are limited in that they often focus on simple metrics, end-to-end outcomes, and are dependent on the perspectives of humans. In order to match the increasing complexity of these many component systems, evaluation frameworks must also be able to reason, probe, iterate, and understand the nuanced logic passing through these systems. In this paper, we present RAFFLES, an evaluation architecture that incorporates reasoning and iterative refinement. Specifically, RAFFLES operates as an iterative, multi-component pipeline, using a central Judge to systematically investigate faults and a set of specialized Evaluators to assess not only the system's components but also the quality of the reasoning by the Judge itself. We evaluated RAFFLES with several benchmarks - the Who&When dataset to identify step-level faults in agentic systems and the ReasonEval datasets to diagnose step-level mathematical reasoning errors. RAFFLES outperforms strong baselines, achieving an accuracy of over 20% and 50% on the Who&When Hand-Crafted and Algorithm datasets, and over 80% on the ReasonEval datasets. These results demonstrate a key step towards introducing automated fault detection for autonomous systems over labor-intensive manual human review.

RAFFLES: Reasoning-based Attribution of Faults for LLM Systems

Large vision-language models (LVLMs) have demonstrated outstanding performance in many downstream tasks. However, LVLMs are trained on large-scale datasets, which can pose privacy risks if training images contain sensitive information. Therefore, it is important to detect whether an image is used to train the LVLM. Recent studies have investigated membership inference attacks (MIAs) against LVLMs, including detecting image-text pairs and single-modality content. In this work, we focus on detecting whether a target image is used to train the target LVLM. We design simple yet effective Image Corruption-Inspired Membership Inference Attacks (ICIMIA) against LLVLMs, which are inspired by LVLM's different sensitivity to image corruption for member and non-member images. We first perform an MIA method under the white-box setting, where we can obtain the embeddings of the image through the vision part of the target LVLM. The attacks are based on embedding similarity between the image and its corrupted version. We further explore a more practical scenario where we have no knowledge about target LVLMs and we can only query the target LVLMs with an image and a question. We then conduct the attack by utilizing the output text embeddings' similarity. Experiments on existing datasets validate the effectiveness of our proposed attack methods under two different settings.

Image Corruption-Inspired Membership Inference Attacks against Large Vision-Language Models

Specialized visual tools can augment large language models or vision language models with expert knowledge (e.g., grounding, spatial reasoning, medical knowledge, etc.), but knowing which tools to call (and when to call them) can be challenging. We introduce DART, a multi-agent framework that uses disagreements between multiple debating visual agents to identify useful visual tools (e.g., object detection, OCR, spatial reasoning, etc.) that can resolve inter-agent disagreement. These tools allow for fruitful multi-agent discussion by introducing new information, and by providing tool-aligned agreement scores that highlight agents in agreement with expert tools, thereby facilitating discussion. We utilize an aggregator agent to select the best answer by providing the agent outputs and tool information. We test DART on four diverse benchmarks and show that our approach improves over multi-agent debate as well as over single agent tool-calling frameworks, beating the next-strongest baseline (multi-agent debate with a judge model) by 3.4% and 2.4% on A-OKVQA and MMMU respectively. We also find that DART adapts well to new tools in applied domains, with a 1.3% improvement on the M3D medical dataset over other strong tool-calling, single agent, and multi-agent baselines. Additionally, we measure text overlap across rounds to highlight the rich discussion in DART compared to existing multi-agent methods. Finally, we study the distribution of expert tool calls to ensure that every tool is being reliably used to help resolve disagreement.

DART: Leveraging Multi-Agent Disagreement for Tool Recruitment in Multimodal Reasoning

Reliable mathematical and scientific reasoning remains an open challenge for large visionвЂ“language models (VLMs). Standard final-answer evaluation often masks reasoning errors, allowing silent failures to persist. To address this gap, we introduce TRACE (Transparent Reasoning And Consistency Evaluation), a framework for analyzing, diagnosing, and improving reasoning in VLMs. At its core, TRACE leverages Auxiliary Reasoning Sets (ARS), compact sub-questionвЂ“answer pairs that decompose complex problems, evaluate intermediate steps through consistency-based metrics, and expose failures overlooked by standard evaluation. Our experiments show that consistency across ARS is linked to final-answer correctness and helps pinpoint the reasoning steps where failures arise, offering actionable signals for model improvement.

TRACE: A Framework for Analyzing and Enhancing Stepwise Reasoning in Vision-Language Models

Understanding how large language models (LLMs) process compositional linguistic structures is integral to enhancing their reliability and interpretability. We present Constituent-Aware Pooling (CAP), a methodology grounded in compositionality, mechanistic interpretability, and information theory that intervenes in model activations by pooling token representations into linguistic constituents at various layers. Experiments across eight models (124M-8B parameters) on inverse definition modelling, hypernym and synonym prediction reveal that semantic composition is not localised to specific layers but distributed across network depth. Performance degrades substantially under constituent-based pooling, particularly in early and middle layers, with larger models showing greater sensitivity. We propose an information-theoretic interpretation: transformers' training objectives incentivise deferred integration to maximise token-level throughput, resulting in fragmented rather than localised composition. These findings highlight fundamental architectural and training constraints requiring specialised approaches to encourage robust compositional processing.

Where Do LLMs Compose Meaning? A Layerwise Analysis of Compositional Robustness

We investigate how the choice of output format influences the fine-tuning performance of large language models on sequence labelling- based information extraction tasks. Across more than 280 experiments spanning multiple benchmarks, model families, and formats, we find that the output format is a critical yet often overlooked hyperparameter. In some cases, the choice of format alone can shift F1 scores by over 40% on the same dataset and model. Further, we observe that no single format consistently dominates across settings - the optimal choice depends on factors like model family, size and dataset characteristics. Overall, these results demonstrate that even informationally equivalent output formats can produce substantial performance variation, highlighting the need to treat output formatting as a key factor in building robust and reliable information extraction systems.

Lost in Formatting: How Output Formats Skew LLM Performance on Information Extraction

Recent advances in mathematical reasoning typically rely on massive scale, yet the question remains: can strong reasoning capabilities be induced in small language models (<=1.5B) under extreme constraints? We investigate this by training models on a single A40 GPU (48GB) for under 24 hours using Reinforcement Learning with Verifiable Rewards (RLVR) and Low-Rank Adaptation (LoRA). We find that the success of this ``micro-budget" regime depends critically on the interplay between adapter capacity and model initialization. While low-rank adapters (r=8) consistently fail to capture the complex optimization dynamics of reasoning, high-rank adapters (r=256) unlock significant plasticity in standard instruction-tuned models. Our best result achieved an impressive 40.0% Pass@1 on AIME 24 (an 11.1% absolute improvement over baseline) and pushed Pass@16 to 70.0%, demonstrating robust exploration capabilities. However, this plasticity is not universal: while instruction-tuned models utilized the budget to elongate their chain-of-thought and maximize reward, heavily math-aligned models suffered performance collapse, suggesting that noisy, low-budget RL updates can act as destructive interference for models already residing near a task-specific optimum.

Plasticity vs. Rigidity: The Impact of Low-Rank Adapters on Reasoning on a Micro-Budget

The rapid adoption of Small Language Models (SLMs) for resource constrained applications has outpaced our understanding of their ethical and fairness implications. To address this gap, we introduce the Vacuous Neutrality Framework (VaNeu), a multi-dimensional evaluation paradigm designed to assess SLM fairness prior to deployment. The framework examines model robustness across four stages - biases, utility, ambiguity handling, and positional bias over diverse social bias categories. To the best of our knowledge, this work presents the first large-scale audit of SLMs in the 0.5–5B parameter range, an overlooked “middle tier” between BERT-class encoders and flagship LLMs. We evaluate nine widely used SLMs spanning four model families under both ambiguous and disambiguated contexts. Our findings show that models demonstrating low bias in early stages often fail subsequent evaluations, revealing hidden vulnerabilities and unreliable reasoning. These results underscore the need for a more comprehensive understanding of fairness and reliability in SLMs, and position the proposed framework as a principled tool for responsible deployment in socially sensitive settings. The code is available at: https://github.com/smanduru10/Vacuous-Neutrality-Framework.git.

Beyond Bias Scores: Unmasking Vacuous Neutrality in Small Language Models

Retrieval-augmented generation has become the dominant paradigm for deploying large language models in knowledge-intensive applications, yet practitioners lack guidance on model selection when both quality and costs matter. We evaluate language models from 4B to 70B parameters, including PLLuM and Bielik families of Polish LLM, within a Polish Wikipedia-based RAG pipeline. Quality assessment uses GPT-4o pairwise comparison across 1,000 PolQA questions with bias mitigation and Bradley-Terry ranking, while energy measurements capture inference costs on NVIDIA H100 hardware. Our findings challenge conventional scaling assumptions: parameter scaling beyond 12B offers minimal quality gains, with mid-size PLLuM-12 matching 70B performance while reducing energy consumption by 83\%.

Evaluating Cost-Efficiency of LLMs in a RAG Setup on Polish Wikipedia: Quality vs. Energy Consumption

The ability of AI systems to not only answer complex natural language questions, but also transparently justify their reasoning, is crucial for building trust and enabling effective human-AI collaboration. In domains requiring multi-hop reasoning, answers must often be constructed by combining multiple relevant sentences from a knowledge base to build an inferential path from the question toward the answer. We tackle this challenge by exploring a neuro-symbolic approach to reasoning through the generation of entailment trees -- structured, step-by-step proof trees -- using Large Language Models (LLMs). These trees provide interpretable justifications for the inference process. Using the EntailmentBank~\citep{dalvi2021explaining} data set, we evaluated a diverse set of prompting strategies across multiple models, along with a proposal of an inference-guided prompting approach that performs well. We also fine-tuned LLMs trained specifically for proof generation by applying several data augmentation, curriculum learning, and reinforcement-guided optimization strategies. Our results show that the fine-tuned model outperforms all prompting strategies, achieving superior performance across multiple structural and semantic metrics. We also provide a detailed evaluation of which training strategies are helpful towards proof generation. Our findings highlight the importance of proof tree generation as a benchmark for evaluating structured reasoning in LLMs.

Downloads

Next from EACL 2026

RAFFLES: Reasoning-based Attribution of Faults for LLM Systems

Similar lecture

Emergence of Episodic Memory in Transformers: Characterizing Changes in Temporal Structure of Attention Scores During Training

.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

RAFFLES: Reasoning-based Attribution of Faults for LLM Systems

Similar lecture

Emergence of Episodic Memory in Transformers: Characterizing Changes in Temporal Structure of Attention Scores During Training

Downloads