United States

LLMs are crucial for generative AI, but their causal reasoning abilities remain uncertain, especially at Pearl’s interventional and counterfactual levels. This study presents a dataset of over 7,000 causal questions in Portuguese, sourced from real human interactions. A subset was annotated using a novel taxonomy based on causal intent, action requirements, and the required level of reasoning. Evaluations with GPT-4o and Llama3.1 reveal significant challenges in advanced causal tasks, highlighting LLMs’ limitations and the need for improvements in causal reasoning for AI agents.

NAACL 2025

Advancing Causal Reasoning in Portuguese: An Evaluation of LLMs in Interventions and Counterfactuals

interventional reasoning

causal reasoning

large language models

counterfactual reasoning

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. 
[Full Message from General Chair Philipp Koehn](https://drive.google.com/file/d/1QiJFOmMjvDgsDo3i1XzlQnImxe3RU_Cy/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://drive.google.com/file/d/1QiJFOmMjvDgsDo3i1XzlQnImxe3RU_Cy/view?usp=sharing')">Conference Handbook</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 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.

**Online Registration Form**: https://acl.swoogo.com/acl2026

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.

poster

### 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!

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

Large Language Models (LLMs) have excelled in multi-hop question-answering (M-QA) due to their advanced reasoning abilities. However, the impact of inherent reasoning structures on LLM M-QA performance remains unclear. To fill this gap, we introduce the Graph Reasoning-Structured Question Answering Dataset (GRS-QA), which includes both semantic contexts and reasoning structures for QA pairs. Unlike existing M-QA~datasets, where different reasoning structures are entangled, GRS-QA explicitly captures reasoning pathways through reasoning graphs, where nodes represent textual contexts and edges denote logical flows. These graphs enable fine-grained evaluation of LLM reasoning across various structures. Our empirical analysis shows that LLMs perform differently based on reasoning structures, highlighting the role of textual structure in reasoning beyond semantics.

GRS-QA - Graph Reasoning-Structured Question Answering Dataset

Generating cell-level descriptions for Jupyter Notebooks, which is a major resource consisting of codes, tables, and descriptions, has been attracting increasing research attention. However, existing methods for Jupyter Notebooks mostly focus on generating descriptions from code snippets or table outputs independently. On the other side, descriptions should be personalized as users have different purposes in different scenarios while previous work ignored this situation during description generation. In this work, we formulate a new task, personalized description generation with code, tables, and user-written guidelines in Jupyter Notebooks. To evaluate this new task, we collect and propose a benchmark, namely TOOL , containing code, tables, and user-written guidelines as inputs and personalized descriptions as targets. Extensive experiments show that while existing models of text generation are able to generate fluent and readable descriptions, they struggle to produce factually correct descriptions without user-written guidelines based on automatic and human evaluation. Common error patterns include guideline misalignment, incorrect variables, missing code details, and faulty reasoning. Ablations show that implementing guidelines substantially improves performance. Our downstream application examining Human-AI collaboration in the documentation for code-table demonstrated enhanced usability and effectiveness.

NBDESCRIB: A Dataset for Text Description Generation from Tables and Code in Jupyter Notebooks with Guidelines

Recent advances in video-text retrieval (VTR) have largely relied on supervised learning and fine-tuning. In this paper, we introduce , a novel zero-shot VTR framework that leverages off-the-shelf video captioners, large language models (LLMs), and text retrieval methods—entirely without additional training or annotated data. Due to the limited power of captioning methods, the captions often miss important content in the video, resulting in unsatisfactory retrieval performance. To translate more information into video captions, we first generates initial captions for videos, then enhances them using a relevance-boosted captioning strategy powered by LLMs, enriching video descriptions with salient details. To further emphasize key content, we propose structural information extraction, organizing visual elements such as objects, events, and attributes into structured templates, further boosting the retrieval performance. Benefiting from the enriched captions and structuralized information, extensive experiments on several video-text retrieval benchmarks demonstrate the superiority of over existing fine-tuned and pretraining methods without any data. They also show that the enriched captions capture key details from the video with minimal noise. Code and data will be released to facilitate future research.

ELIOT: Zero-Shot Video-Text Retrieval through Relevance-Boosted Captioning and Structural Information Extraction

Commonsense inference and domain-specific expertise are crucial for understanding and responding to emotional, cognitive, and topic-specific cues in counseling conversations with crime victims. However, these key evidences are often dispersed across multiple utterances, making it difficult to capture through single-hop reasoning. To address this, we propose MENDER, a novel Multi-hop commonsensE and domaiN-specific Chain-of-Thought (CoT) reasoning framework for knowleDge-grounded empathEtic Response generation in counseling dialogues. MENDER leverages large language models (LLMs) to integrate commonsense and domain knowledge via multi-hop reasoning over the dialogue context. It employs two specialized reasoning chains, viz. Commonsense Knowledge-driven CoT and Domain Knowledge-driven CoT rationales, which extract and aggregate dispersed emotional, cognitive, and topical evidences to generate knowledge-grounded empathetic counseling responses. Experimental evaluations on counseling dialogue dataset, POEM validate MENDER’s efficacy in generating coherent, empathetic, knowledge-grounded responses.

MENDER: Multi-hop Commonsense and Domain-specific CoT Reasoning for Knowledge-grounded Empathetic Counseling of Crime Victims

Large language models (LLMs) have gained popularity in recent years for their utility in various applications. However, they are sensitive to non-semantic changes in prompt formats, where small changes in the prompt format can lead to significant performance fluctuations. In the literature, this problem is commonly referred to as prompt brittleness. Previous research on prompt engineering has focused mainly on developing techniques for identifying the optimal prompt for specific tasks. Some studies have also explored the issue of prompt brittleness and proposed methods to quantify performance variations; however, no simple solution has been found to address this challenge. We propose Mixture of Formats (MOF), a simple and efficient technique for addressing prompt brittleness in LLMs by diversifying the styles used in the prompt few-shot examples. MOF was inspired by computer vision techniques that utilize diverse style datasets to prevent models from associating specific styles with the target variable. Empirical results show that our proposed technique reduces style-induced prompt brittleness in various LLMs while also enhancing overall performance across prompt variations and different datasets.

Towards LLMs Robustness to Changes in Prompt Format Styles

We analyze contextual representations in neural autoregressive language models, emphasizing long-range contexts that span several thousand tokens. Our methodology employs a perturbation setup and the metric Anisotropy-Calibrated Cosine Similarity, to capture the degree of contextualization of long-range patterns from the perspective of representation geometry. We begin the analysis with a case study on standard decoder-only Transformers, demonstrating that similar perplexity can exhibit markedly different downstream task performance, which can be explained by the difference in contextualization of long-range content. Next, we extend the analysis to other models, covering recent novel architectural designs and various training configurations. The representation-level results illustrate a reduced capacity for high-complexity (i.e., less compressible) sequences across architectures, and that fully recurrent models rely heavily on local context, whereas hybrid models more effectively encode the entire sequence structure. Finally, preliminary analysis of model size and training configurations on the encoding of long-range context suggest potential directions for improving existing language models.

How much do contextualized representations encode long-range context?

Retrieval-Augmented Generation (RAG) models have drawn considerable attention in modern open-domain question answering. The effectiveness of RAG depends on the quality of the top retrieved documents. However, conventional retrieval methods sometimes fail to rank the most relevant documents at the top. In this paper, we introduce ASRANK, a new re-ranking method based on scoring retrieved documents using zero-shot answer scent which relies on a pre-trained large language model to compute the likelihood of the document-derived answers aligning with the answer scent. Our approach demonstrates marked improvements across several datasets, including NQ, TriviaQA, WebQA, ArchivalQA, HotpotQA, and Entity Questions. Notably, ASRANK increases Top-1 retrieval accuracy on NQ from 19.2% to 46.5% for MSS and 22.1% to 47.3% for BM25. It also shows strong retrieval performance on several datasets compared to state-of-the-art methods (47.3 Top-1 by ASRANK vs 35.4 by UPR by BM25).

ASRank: Zero-Shot Re-Ranking with Answer Scent for Document Retrieval

As safety remains a crucial concern throughout the development lifecycle of Large Language Models (LLMs), researchers and industrial practitioners have increasingly focused on safeguarding and aligning LLM behaviors with human preferences and ethical standards. LLMs, trained on extensive multilingual corpora, exhibit powerful generalization abilities across diverse languages and domains. However, current safety alignment practices predominantly focus on single-language scenarios, which leaves their effectiveness in complex multilingual contexts, especially for those complex mixed-language formats, largely unexplored. In this study, we introduce Multilingual Blending, a mixed-language query-response scheme designed to evaluate the safety alignment of various state-of-the-art LLMs (e.g., GPT-4o, GPT 3.5, Llama3) under sophisticated, multilingual conditions. We further investigate language patterns such as language availability, morphology, and language family that could impact the effectiveness of Multilingual Blending in compromising the safeguards of LLMs. Our experimental results show that, without meticulously crafted prompt templates, Multilingual Blending significantly amplifies the detriment of malicious queries, leading to dramatically increased bypass rates in LLM safety alignment (67.23% on GPT-3.5 and 40.34% on GPT-4o), far exceeding those of single-language baselines. Moreover, the performance of Multilingual Blending varies notably based on intrinsic linguistic properties, with languages of different morphology and from diverse families being more prone to evading safety alignments. These findings underscore the necessity of evaluating LLMs and developing corresponding safety alignment strategies in a complex, multilingual context to align with their superior cross-language generalization capabilities.

Multilingual Blending: Large Language Model Safety Alignment Evaluation with Language Mixture

Large Language Models (LLMs) have recently demonstrated remarkable performance in Text-to-SQL tasks. However, existing research primarily focuses on the optimization of prompts and improvements in workflow, with few studies delving into the exploration of the questions. In this paper, we propose a Text-to-SQL framework based on question classification and multi-agent collaboration (QCMA-SQL). Specifically, we first employ multiple cross-attention mechanisms to train a schema selector to classify questions and select the most suitable database schema. Subsequently, we employ the appropriate agents based on the varying difficulty levels of the questions to generate preliminary SQL queries. Moreover, we implement syntax validation and execution optimization steps to generate final SQL queries. Experimental results on the Spider dataset show that the QCMA-SQL framework achieves an execution accuracy of 87.4%, outperforming state-of-the-art methods. Through ablation studies, we find that classifying the questions ultimately leads to a 2.8% increase in execution accuracy.

Enhancing Text-to-SQL with Question Classification and Multi-Agent Collaboration

Large language models (LLMs) have seen rapid development in recent years, revolutionizing various applications and significantly enhancing convenience and productivity. However, alongside their impressive capabilities, ethical concerns and new types of attacks, such as jailbreaking, have emerged. While most prompting techniques focus on optimizing adversarial inputs for individual cases, resulting in higher computational costs when dealing with large datasets. Less research has addressed the more general setting of training a universal attacker that can transfer to unseen tasks. In this paper, we introduce JUMP, a prompt-based method designed to jailbreak LLMs using universal multi-prompts. We also adapt our approach for defense, which we term DUMP. Experimental results demonstrate that our method for optimizing universal multi-prompts outperforms existing techniques.

Downloads

Next from NAACL 2025

GRS-QA - Graph Reasoning-Structured Question Answering Dataset

Similar lecture

Causal-Audit: Explicit and Auditable Graph-based Reasoning via Target-Aware Causal Chain Construction