Deep neural networks (DNN) have achieved unprecedented success in numerous machine
learning tasks in various domains. However, the existence of adversarial examples raises …

Language Models (LMs) often cannot be deployed because of their potential to harm users
in hard-to-predict ways. Prior work identifies harmful behaviors before deployment by using …

Adversarial examples highlight model vulnerabilities and are useful for evaluation and
interpretation. We define universal adversarial triggers: input-agnostic sequences of tokens …

Chatbots are used in many applications, eg, automated agents, smart home assistants,
interactive characters in online games, etc. Therefore, it is crucial to ensure they do not …

Open-domain dialog generation is a challenging problem; maximum likelihood training can
lead to repetitive outputs, models have difficulty tracking long-term conversational goals, and …

Although deep learning models have brought tremendous advancements to the field of open-
domain dialogue response generation, recent research results have revealed that the …

Text infilling is defined as a task for filling in the missing part of a sentence or paragraph,
which is suitable for many real-world natural language generation scenarios. However …

Large pretrained language models can easily produce toxic or biased content, which is
prohibitive for practical use. In order to detect such toxic generations, existing methods rely …

In recent years there is a surge of interest in the interpretability and explainability of AI
systems, which is largely motivated by the need for ensuring the transparency and …

Neural dialogue models have been widely adopted in various chatbot applications because
of their good performance in simulating and generalizing human conversations. However …