This paper focuses on the task of word sense disambiguation (WSD) on lexicographic examples relying on the French Lexical Network (fr-LN). For this purpose, we exploit the lexical and relational properties of the netw...
详细信息
In psychology and neuroscience, dreams are extensively studied both as a model to understand the neural bases of consciousness and for their relationship with psycho-physical well-being. The study of dream content typ...
详细信息
Rooted in AMR, Uniform Meaning Representation (UMR) is a graph-based formalism with nodes as concepts and edges as relations between them. When used to represent naturallanguage semantics, UMR maps words in a sentenc...
详细信息
Large language models (LLMs) have demonstrated exceptional capabilities in naturallanguageprocessing tasks, fueling innovations in emerging areas such as the metaverse. These models enable dynamic virtual communitie...
详细信息
ISBN:
(纸本)9798400712999
Large language models (LLMs) have demonstrated exceptional capabilities in naturallanguageprocessing tasks, fueling innovations in emerging areas such as the metaverse. These models enable dynamic virtual communities, enhancing user interactions and revolutionizing industries. However, their increasing deployment exposes vulnerabilities to jailbreak attacks, where adversaries can manipulate LLMdriven systems to generate harmful content. While various defense mechanisms have been proposed, their efficacy against diverse jailbreak techniques remains unclear. This paper addresses this gap by evaluating the performance of three popular defense methods (Backtranslation, Self-reminder, and Paraphrase) against different jailbreak attack strategies (GCG, BEAST, and Deepinception), while also utilizing three distinct models. Our findings reveal that while defenses are highly effective against optimization-based jailbreak attacks and reduce the attack success rate by 79% on average, they struggle in defending against attacks that alter attack motivations. Additionally, methods relying on self-reminding perform better when integrated with models featuring robust safety guardrails. For instance, Llama27b shows a 100% reduction in Attack Success Rate, while Vicuna-7b and Mistral-7b, lacking safety alignment, exhibit a lower average reduction of 65.8%. This study highlights the challenges in developing universal defense solutions for securing LLMs in dynamic environments like the metaverse. Furthermore, our study highlights that the three distinct models utilized demonstrate varying initial defense performance against different jailbreak attack strategies, underscoring the complexity of effectively securing LLMs.
This paper presents the MEDIQA-Chat 2023 shared task organized at the ACL-Clinical NLP workshop. The shared task is motivated by the need to develop methods to automatically generate clinical notes from doctor-patient...
详细信息
Knowledge graphs provide significant assistance for many artificial intelligence tasks, but they are usually incomplete. Techniques for knowledge graph completion can improve the coverage of Knowledge graphs (KGs) by ...
详细信息
The advent of large language models has brought about new ways of interacting with data intuitively via naturallanguage. In recent years, a variety of visualization systems have explored the use of naturallanguage t...
详细信息
This paper summarizes two approaches developed for BioNLP2023 workshop task 1A 1: clinical progress note summarization. We develop two types of methods with either rules or pre-trained language models. In the rule-bas...
详细信息
Multiword expressions are a key ingredient for developing large-scale and linguistically sound naturallanguageprocessing technology. This paper describes our improvements in automatically identifying Romanian multiw...
详细信息
In this work, we propose the application of abstract meaning representation (AMR) based semantic parsing models to parse textual descriptions of a visual scene into scene graphs, which is the first work to the best of...
详细信息
暂无评论