United States

Gradient optimization-based adversarial attack methods automate the learning of adversarial triggers to generate jailbreak prompts or leak system prompts. In this work, we take a closer look at the optimization objective of adversarial trigger learning and propose ATLA: Adversarial Trigger Learning with Augmented objectives. ATLA improves the negative log-likelihood loss used by previous studies into a weighted loss formulation that encourages the learned adversarial triggers to optimize more towards response format tokens. This enables ATLA to learn an adversarial trigger from just one query-response pair and the learned trigger generalizes well to other similar queries. We further design a variation to augment trigger optimization with an auxiliary loss that suppresses evasive responses. We showcase how to use ATLA to learn adversarial suffixes jailbreaking LLMs and to extract hidden system prompts. Empirically we demonstrate that ATLA consistently outperforms current state-of-the-art techniques, achieving nearly 100% success in attacking while requiring 80% fewer queries. ATLA learned jailbreak suffixes demonstrate high generalization to unseen queries and transfer well to new LLMs.

NAACL 2025

Augmented Adversarial Trigger Learning

jailbreaking

red-teaming

llms

### Welcome to The 63rd Annual Meeting of the Association for Computational Linguistics (ACL 2025)

Message from the General Chair: 
*It is my great pleasure and honor to welcome you to the 63rd Annual Meeting of the Association for Computational Linguistics (ACL 2025), held in beautiful Vienna, Austria, from July 27 to August 1, 2025. ACL2025continues our field’s 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.* 
[Read more](https://drive.google.com/file/d/1GI_hvOpjswAuYdUTromfeDiPpCcqidwg/view?usp=sharing)

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 2025". 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 need to log in with the **email address you registered with**. 

ACL 2025

Welcome to The 63rd Annual Meeting of the Association for Computational Linguistics (ACL 2025)

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

Symbolic music is represented in two distinct forms: two-dimensional, visually intuitive score images, and one-dimensional, standardized text annotation sequences. While large language models have shown extraordinary potential in music, current research has primarily focused on unimodal symbol sequence text. Existing general-domain visual language models still lack the ability of music notation understanding. Recognizing this gap, we propose NOTA, the first large-scale comprehensive multimodal music notation dataset. It consists of 1,019,237 records, from 3 regions of the world, and contains 3 tasks. Based on the dataset, we trained NotaGPT, a music notation visual large language model. Specifically, we involve a pre-alignment training phase for cross-modal alignment between the musical notes depicted in music score images and their textual representation in ABC notation. Subsequent training phases focus on foundational music information extraction, followed by training on music score notation analysis. Experimental results demonstrate that our NotaGPT-7B achieves significant improvement on music understanding, showcasing the effectiveness of NOTA and the training pipeline.

NOTA: Multimodal Music Notation Understanding for Visual Large Language Model

For centuries, writers have hidden messages as acrostics, in which initial letters of consecutive lines or paragraphs form meaningful words or phrases. Scholars searching for acrostics manually can only focus on a few authors at a time and often favor qualitative arguments about whether a given acrostic is accidental or intentional. Here we describe AcrosticSleuth, a first-of-its-kind approach to identify acrostics automatically and rank them by the probability that the corresponding sequence of characters does not occur by chance. Since acrostics are rare, we formalize the problem as a binary classification task in the presence of extreme class imbalance. To evaluate AcrosticSleuth, we present the Acrostic Identification Dataset (AcrostID), a collection of acrostics from the WikiSource online database. Despite the class imbalance, AcrosticSleuth achieves F1 scores of 0.39, 0.59, and 0.66 on the French, English, and Russian subdomains of WikiSource, respectively. We further demonstrate that AcrosticSleuth can identify previously unknown instances of wordplay in high-profile literary contexts, including the English philosopher Thomas Hobbes' signature in the opening paragraphs of The Elements of Law.

AcrosticSleuth: Probabilistic Identification and Ranking of Acrostics in Multilingual Corpora

Parameter-efficient fine-tuning techniques like Low-Rank Adaptation (LoRA) have revolutionized the adaptation of large language models (LLMs) to diverse tasks. Recent efforts have explored mixtures of LoRA modules for multi-task settings. However, our analysis reveals redundancy in the down-projection matrices of these architectures. This observation motivates our proposed method, Mixture of Dyadic Experts (MoDE), which introduces a novel design for efficient multi-task adaptation. This is done by sharing the down-projection matrix across tasks and employing atomic rank-one adapters, coupled with routers that allow more sophisticated task-level specialization. Our design allows for more fine-grained mixing, thereby increasing the model's ability to jointly handle multiple tasks. We evaluate MoDE on the Supernatural Instructions (SNI) benchmark consisting of a diverse set of 700+ tasks and demonstrate that it outperforms state-of-the-art multi-task parameter-efficient fine-tuning (PEFT) methods, without introducing additional parameters. Our findings contribute to a deeper understanding of parameter efficiency in multi-task LLM adaptation and provide a practical solution for deploying high-performing, lightweight models.

MoDE: Effective Multi-task Parameter Efficient Fine-Tuning with a Mixture of Dyadic Experts

Speculative Decoding (SD) utilizes an efficient draft model to generate multiple tokens, which are subsequently verified in parallel by a target model. This approach has shown significant potential for accelerating inference in large language models (LLMs), with performance heavily reliant on the hyperparameter K—the window size. However, previous methods often depend on simple heuristics to select K or dynamically adjust the window size, which may necessitate additional training or careful resource management to avoid competition. To address these challenges, we propose **H**ierarchical **S**peculative **D**ecoding with **D**ynamic **W**indow (HSDDW), a straightforward framework that eliminates the need for additional training. Specifically, we introduce a *self-verify* mechanism that enables the draft model to autonomously decide when to stop generating tokens. Additionally, by integrating a hierarchical structure that leverages the capabilities of models of different sizes, we significantly enhance the overall speed of the system. HSDDW demonstrates competitive performance across four datasets, achieving notable speedups of 2.91× on MT-Bench and 2.99× on Alpaca, outperforming existing state-of-the-art methods.

Hierarchical Speculative Decoding with Dynamic Window

To tackle the issue of driver fatigue, we propose a novel learning-based multi-frame visual feature framework (LMVFF) for fatigue detection. The framework follows these steps: First, we detect face landmarks and calculate the distances between the eyes, lips, and the head rotation angle based on 68 landmarks. Next, we extract features from the eye region and develop a visual model that classifies eye openness. We then combine the lip region’s open and close features to detect yawning further and assess driver fatigue by monitoring the frequency of eye blinks, yawns, and head movements. LMVFF significantly reduces fatigue mis-identifications compared to single-feature detection methods. By integrating three decision-making strategies, our approach achieves high accuracy in fatigue detection. Additionally, we apply various quantization and compression techniques to models for face detection, landmark recognition, and eye and lip region analysis, greatly enhancing detection speed. The system operates at 25-30 FPS, meeting the real-time performance requirements.

A Learning-based Multi-Frame Visual Feature Framework for Real-Time Driver Fatigue Detection

Recent studies highlight the reliance of Large Language Models (LLMs) on high-quality, diverse data for optimal performance. The data sourced from the Internet often aggregated into datasets like the Common Crawl corpus, presents significant quality variability and necessitates extensive cleaning. Moreover, specific domain knowledge is usually presented in HTML, but there is a lack of effective methods to clean them into the training corpus automatically. Traditional cleaning methods involve either labor-intensive human teams that lack scalability or static heuristics that lead to suboptimal outcomes and are unable to be applied to specific target domains. In this paper, inspired by the recent progress in employing LLMs as versatile agents for diverse tasks, we take the initiative to explore the potential of these agents in automating data-cleaning methodologies. By configuring LLMs as an agent team that imitates the human data-cleaning team, we can automatically generate cleaning rules that traditionally require the involvement of data-cleaning experts. These rules are developed using a limited number of data samples and can then be applied broadly to substantial portions of raw data from the same domain. We demonstrate the efficiency and effectiveness of OURS on both pre-train scale corpora such as Common Crawl and specific target websites. Both automatic and human evaluations of the quality of the cleaned content highlight the feasibility of using LLMs to prepare their training corpus.

AutoClean: LLMs Can Prepare Their Training Corpus

With the rapid advancement of large language models (LLMs), recent years have witnessed many promising studies on leveraging LLM-based agents to simulate human social behavior. While prior work has demonstrated significant potential across various domains, much of it has focused on specific scenarios involving a limited number of agents and has lacked the ability to adapt when errors occur during simulation. To overcome these limitations, we propose a novel LLM-agent-based simulation platform called GenSim, which: (1) Abstracts a set of general functions to simplify the simulation of customized social scenarios; (2) Supports one hundred thousand agents to better simulate large-scale populations in real-world contexts; (3) Incorporates error-correction mechanisms to ensure more reliable and long-term simulations. To evaluate our platform, we assess both the efficiency of large-scale agent simulations and the effectiveness of the error-correction mechanisms. To our knowledge, GenSim represents an initial step toward a general, large-scale, and correctable social simulation platform based on LLM agents, promising to further advance the field of social science.

GenSim: A General Social Simulation Platform with Large Language Model based Agents

The rapid advancement of large language models (LLMs) and artificial intelligence-generated content (AIGC) has accelerated AI-native applications, such as AI-based storybooks that automate engaging story production for children. However, challenges remain in improving story attractiveness, enriching storytelling expressiveness, and developing open-source evaluation benchmarks and frameworks. Therefore, we propose and opensource MM-StoryAgent, which creates immersive narrated video storybooks with refined plots, role-consistent images, and multi-channel audio. MM-StoryAgent designs a multi-agent framework that employs LLMs and diverse expert tools (generative models and APIs) across several modalities to produce expressive storytelling videos. The framework enhances story attractiveness through a multi-stage writing pipeline. In addition, it improves the immersive storytelling experience by integrating sound effects with visual, music and narrative assets. MM-StoryAgent offers a flexible, open-source platform for further development, where generative modules can be substituted. Both objective and subjective evaluation regarding textual story quality and alignment between modalities validate the effectiveness of our proposed MM-StoryAgent system. The demo and source code are available.

MM-StoryAgent: Immersive Narrated Storybook Video Generation with a Multi-Agent Paradigm across Text, Image and Audio

LangLearn is an open-source framework designed to facilitate autonomous learning of low-resource languages (LRL). By combining a language-agnostic approach with AI-enhanced flashcards, LangLearn empowers users to generate custom flashcards for their vocabulary, while offering structured learning through both pre-curated and self-curated decks. The framework integrates six key components: the word definition, corresponding Hanji characters, romanization with numeric tones, audio pronunciation, a sample sentence, as well as a contextual AI-generated image. LangLearn currently supports English and Taiwanese Hokkien (a variety of Southern Min), with plans to extend support for other dialects. Our preliminary study demonstrates that LangLearn positively empowers users to engage with LRLs using their vocabulary preferences, with a comprehensive user study currently underway. LangLearn's modular structure enables future expansion, including ASR-based pronunciation practice. The code is available at https://github.com/HokkienTranslation/HokkienTranslation.

Learning Low-Resource Languages Through NLP-Driven Flashcards: A Case Study of Hokkien in Language Learning Applications

In recent research advancements within the community, large language models (LLMs) have sparked great interest in creating autonomous agents. However, current prompt-based agents often heavily rely on large-scale LLMs. Meanwhile, although fine-tuning methods significantly enhance the capabilities of smaller LLMs, the fine-tuned agents often lack the potential for self-reflection and self-improvement. To address these challenges, we introduce a novel agent framework named RetroAct, which is a framework that jointly optimizes both task-planning and self-reflective evolution capabilities in language agents. Specifically, we develop a two-stage joint optimization process that integrates imitation learning and reinforcement learning, and design an off-policy joint policy gradient optimization algorithm with imitation learning regularization to enhance the data efficiency and training stability in agent tasks. RetroAct significantly improves the performance of open-source models, reduces dependency on closed-source LLMs, and enables fine-tuned agents to learn and evolve continuously. We conduct extensive experiments across various testing environments, demonstrating RetroAct has substantial improvements in task performance and decision-making processes.

Downloads

Next from NAACL 2025

NOTA: Multimodal Music Notation Understanding for Visual Large Language Model

Similar lecture

Investigating Adversarial Trigger Transfer in Large Language Models