Mexico

Large Language Models&#39; safety remains a critical concern due to their vulnerability to jailbreaking attacks, which can prompt these systems to produce harmful and malicious responses. Safety classifiers, computational models trained to discern and mitigate potentially harmful, offensive, or unethical outputs, offer a practical solution to address this issue. However, despite their potential, existing safety classifiers often fail when exposed to adversarial attacks such as gradient-optimized suffix attacks. In response, our study introduces Adversarial Prompt Shield (APS), a lightweight safety classifier model that excels in detection accuracy and demonstrates resilience against unseen jailbreaking prompts. We also introduce efficiently generated adversarial training datasets, named Bot Adversarial Noisy Dialogue (BAND), which are designed to fortify the classifier&#39;s robustness. Through extensive testing on various safety tasks and unseen jailbreaking attacks, we demonstrate the effectiveness and resilience of our models. Evaluations show that our classifier has the potential to significantly reduce the Attack Success Rate by up to 44.9%. This advance paves the way for the next generation of more reliable and resilient Large Language Models.

NAACL 2024

Robust Safety Classifier Against Jailbreaking Attacks: Adversarial Prompt Shield

safety classifier

llm attacks

adversarial training

adversarial robustness

## Welcome to 2024 Annual Conference of the North American Chapter of the Association for Computational Linguistics!
This year, the conference is in Mexico City. NAACL was actually already planned for Mexico City in 2021, but due to the pandemic the entire conference was moved online. This year, finally, we get to go! So it is my sincere pleasure to welcome you to Mexico City, whether in person or virtually. Having the conference in Mexico City is a good opportunity to emphasize that NAACL is our flagship conference for ACL members not only in North America but also in Central and South America, even though NAACL has been bearing “North” in its name. At this year’s conference, we have a theme to match, with a theme track on the Languages of Latin America to showcase the linguistic diversity of the region.
 
The opportunity to present at NAACL should not depend on a researcher’s travel budget, or their family status. This is why it is so important to make virtual participation at NAACL as good an experience as possible – but we want to also provide a good experience for in-person participants. As a community, we are still working out the best way to do that. This year at NAACL, we are trying out a big virtual poster session ahead of the conference, with the hope that this will make make for a lively and interactive experience. At the same time, we are reducing virtual oral presentations, which seem to be particularly tricky to make to work well. A big thanks to the NAACL program chairs and to Luciana Benotti for all their ideas and work to improve the virtual experience. And participants, virtual as well as in-person: Please let us know what worked for you and what didn’t, so we can continue to improve hybrid conferences. 
 
I have been lucky to work with many amazing people. Without their insight, dedication and patience, and without the many hours of work they put in, NAACL would not have been possible. A huge thank you to the program chairs Helena Gomez, Kevin Duh, and Steve Bethard – you are the best! 
 
Finally, I would like to thank all authors, invited speakers and panelists, area chairs and reviewers, the volunteers organizing and chairing sessions, and all attendees, in-person and virtual. Thank you for helping us make NAACL 2024 come to life. 
 
Welcome and hope you all enjoy the conference! 
Katrin Erk 
The University of Texas at Austin 
NAACL-2024 General Chair 
*You can read the full Welcome message in the [Conference Handbook (downloadable)](https://drive.google.com/file/d/1H1NvW0VASQjkSCYgw3mr-3l4yYDSxz6B/view?usp=sharing)*

To access the event page you need to register [**here.**](http://acl.swoogo.com/naacl2024) 
Your access to the event page is limited based on your registration type. If you registered for workshops only, you will gain access to full workshops content on the day of the workshops program.

Please register!

2024 Annual Conference of the North American Chapter of the Association for Computational Linguistics

Large Language Models' safety remains a critical concern due to their vulnerability to jailbreaking attacks, which can prompt these systems to produce harmful and malicious responses. Safety classifiers, computational models trained to discern and mitigate potentially harmful, offensive, or unethical outputs, offer a practical solution to address this issue. However, despite their potential, existing safety classifiers often fail when exposed to adversarial attacks such as gradient-optimized suffix attacks. In response, our study introduces Adversarial Prompt Shield (APS), a lightweight safety classifier model that excels in detection accuracy and demonstrates resilience against unseen jailbreaking prompts. We also introduce efficiently generated adversarial training datasets, named Bot Adversarial Noisy Dialogue (BAND), which are designed to fortify the classifier's robustness. Through extensive testing on various safety tasks and unseen jailbreaking attacks, we demonstrate the effectiveness and resilience of our models. Evaluations show that our classifier has the potential to significantly reduce the Attack Success Rate by up to 44.9%. This advance paves the way for the next generation of more reliable and resilient Large Language Models.

workshop paper

The creation of pedagogically effective questions is a challenge for teachers and requires significant time and meticulous planning, especially in resource-constrained economies. For example, in India, assessments for social science in high schools are characterized by rote memorization without regard to higher-order skill levels. Automated educational question generation (AEQG) using large language models (LLMs) has the potential to help teachers develop assessments at scale. However, it is important to evaluate the quality and relevance of these questions. In this study, we examine the ability of different LLMs (Falcon 40B, Llama2 70B, Palm 2, GPT 3.5, and GPT 4) to generate relevant and high-quality questions of different cognitive levels, as defined by Bloom's taxonomy. We prompt each model with the same instructions and different contexts to generate 510 questions in the social science curriculum of a state educational board in India. Two human experts used a nine-item rubric to assess linguistic correctness, pedagogical relevance and quality, and adherence to Bloom's skill levels. Our results showed that 91.56% of the LLM-generated questions were relevant and of high quality. This suggests that LLMs can generate relevant and high-quality questions at different cognitive levels, making them useful for creating assessments for scaling education in resource-constrained economies.

How Good are Modern LLMs in Generating Relevant and High-Quality Questions at Different Bloom's Skill Levels for Indian High School Social Science Curriculum?

Cyberbullying has grown in recent years, largely attributed to the proliferation of social media users. This phenomenon manifests in various forms, such as hate speech and offensive language, increasing the necessity of effective detection models to tackle this problem. Most approaches focus on supervised algorithms, which have an important drawback—they heavily depend on the availability of ample training data. This paper attempts to tackle this insufficient data problem using data augmentation (DA) techniques. Concretely, we propose a novel data augmentation technique based on a Diffusion Language Model (DLA). We compare our proposed method against well-known DA techniques, such as contextual augmentation and Easy Data Augmentation (EDA). Our findings reveal a slight but promising improvement, leading to more robust results with very low variance. Additionally, we provide a comprehensive qualitative analysis using classification errors, and complementary analysis, shedding light on the nuances of our approach.

Improving aggressiveness detection using a data augmentation technique based on a Diffusion Language Model

Thanks to the popularity of social media, data generated by online communities provides an abundant source of diverse language information. This abundance of data allows NLP practitioners and computational linguists to analyze sociolinguistic phenomena occurring in digital communication. In this paper, we analyze the Twitter discourse around the Mexican Spanish-speaking LGBT+ community. For this, we evaluate how the polarity of some nouns related to the LGBT+ community has evolved in conversational settings using a corpus of tweets that cover a time span of ten years. We hypothesize that social media's fast-moving, turbulent linguistic environment encourages language evolution faster than ever before. Our results indicate that most of the inspected terms have undergone some shift in denotation or connotation. No other generalizations can be observed in the data, given the difficulty that current NLP methods have to account for polysemy, and the wide differences between the various subgroups that make up the LGBT+ community. A fine-grained analysis of a series of LGBT+-related lexical terms is also included in this work.

The Mexican Gayze: A Computational Analysis of the Attitudes towards the LGBT+ Population in Mexico on Social Media Across a Decade

We present a multi-agent automated fact-checking approach for detecting misinformation by combining an offline LLM agent for claim decomposition with an online web search agent for RAG. Specifically, we demonstrate that using Cohere Coral with a web connector, or DuckDuckGo summarized by GPT-3.5, for web search significantly enhances misinformation detection on the LIAR-New dataset. We evaluate our approach across 5 models, with macro F1 increasing by up to 20 percent.

Web Retrieval Agents for Evidence-Based Misinformation Detection

We investigate the impact of free speech and the relaxation of moderation on online social media platforms using Elon Musk's takeover of Twitter as a case study. By curating a dataset of over 10 million tweets, our study employs a novel framework combining content and network analysis. Our findings reveal a significant increase in the distribution of certain forms of hate content, particularly targeting the LGBTQ+ community and liberals. Network analysis reveals the formation of cohesive hate communities facilitated by influential bridge users, with substantial growth in interactions hinting at increased hate production and diffusion. By tracking the temporal evolution of PageRank, we identify key influencers, primarily self-identified far-right supporters disseminating hate against liberals and woke culture. Ironically, embracing free speech principles appears to have enabled hate speech against the very concept of freedom of expression and free speech itself. Our findings underscore the delicate balance platforms must strike between open expression and robust moderation to curb the proliferation of hate online.

X-posing Free Speech: Examining the Impact of Moderation Relaxation on Online Social Networks

Online gender-based violence has grown concomitantly with the adoption of the internet and social media. Its effects are worse in the Global majority where many users use social media in languages other than English. The scale and volume of conversations on the internet have necessitated the need for automated detection of hate speech and, more specifically, gendered abuse. There is, however, a lack of language-specific and contextual data to build such automated tools. In this paper, we present a dataset on gendered abuse in three languages- Hindi, Tamil and Indian English. The dataset comprises of tweets annotated along three questions pertaining to the experience of gender abuse, by experts who identify as women or a member of the LGBTQIA+ community in South Asia. Through this dataset, we demonstrate a participatory approach to creating datasets that drive AI systems.

The Uli Dataset: An Exercise in Experience Led Annotation of oGBV

Although social media platforms are a prominent arena for users to engage in interpersonal discussions and express opinions, the facade and anonymity offered by social media may allow users to spew hate speech and offensive content. Given the massive scale of such platforms, there arises a need to automatically identify and flag instances of hate speech. Although several hate speech detection methods exist, most of these black-box methods are not interpretable or explainable by design. To address the lack of interpretability, in this paper, we propose to use state-of-the-art Large Language Models (LLMs) to extract features in the form of rationales from the input text, to train a base hate speech classifier, thereby enabling faithful interpretability by design. Our framework effectively combines the textual understanding capabilities of LLMs and the discriminative power of state-of-the-art hate speech classifiers to make these classifiers faithfully interpretable. Our comprehensive evaluation on a variety of social media hate speech datasets demonstrate: (1) the goodness of the LLM-extracted rationales, and (2) the surprising retention of detector performance even after training to ensure interpretability. All code and data will be made available at \url{https://github.com/AmritaBh/shield}.

Towards Interpretable Hate Speech Detection using Large Language Model-extracted Rationales

Natural language processing research has begun to embrace the notion of annotator subjectivity, motivated by variations in labelling. This approach understands each annotator's view as valid, which can be highly suitable for tasks that embed subjectivity, e.g., sentiment analysis. However, this construction may be inappropriate for tasks such as hate speech detection, as it affords equal validity to all positions on e.g., sexism or racism. We argue that the conflation of hate and offence can invalidate findings on hate speech, and call for future work to be situated in theory, disentangling hate from its orthogonal concept, offence.

Subjective Isms? On the Danger of Conflating Hate and Offence in Abusive Language Detection

To address the limitations of current hate speech detection models, we introduce SGHateCheck, a novel framework designed for the linguistic and cultural context of Singapore and Southeast Asia. It extends the functional testing approach of HateCheck and MHC, employing large language models for translation and paraphrasing into Singapore's main languages, and refining these with native annotators. SGHateCheck reveals critical flaws in state-of-the-art models, highlighting their inadequacy in sensitive content moderation. This work aims to foster the development of more effective hate speech detection tools for diverse linguistic environments, particularly for Singapore and Southeast Asia contexts.

SGHateCheck: Functional Tests for Detecting Hate Speech in Low-Resource Languages of Singapore

To make content moderation more culturally sensitive, it is important to create appropriate datasets.
We present AustroTox, a dataset annotated for offensiveness on the token- and post-level.
It comprises 4,562 comments in Austrian German and additionally 1,799 English comments.
AustroTox contains post-level annotations for offensive language and token-level annotations for targets of offensive statements and vulgar language.
An evaluation indicates that GPT-4 performs best in classifying offensiveness on the German part of AustroTox compared to other models. Except for German offensiveness classification, the smaller fine-tuned language models we test excel in classifying offensive comments, detecting vulgar language, and detecting the targets of offensive statements compared to the large language models we prompt in a few-shot fashion.

Downloads

Next from NAACL 2024

How Good are Modern LLMs in Generating Relevant and High-Quality Questions at Different Bloom's Skill Levels for Indian High School Social Science Curriculum?

Similar lecture

Automated Adversarial Discovery for Safety Classifiers

.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 NAACL 2024

How Good are Modern LLMs in Generating Relevant and High-Quality Questions at Different Bloom's Skill Levels for Indian High School Social Science Curriculum?

Similar lecture

Automated Adversarial Discovery for Safety Classifiers

Downloads