@inproceedings{sreedhar-etal-2024-canttalkaboutthis,
    title = "{C}ant{T}alk{A}bout{T}his: Aligning Language Models to Stay on Topic in Dialogues",
    author = "Rebedea, Traian  and
      Sreedhar, Makesh  and
      Ghosh, Shaona  and
      Zeng, Jiaqi  and
      Parisien, Christopher",
    editor = "Al-Onaizan, Yaser  and
      Bansal, Mohit  and
      Chen, Yun-Nung",
    booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2024",
    month = nov,
    year = "2024",
    address = "Miami, Florida, USA",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/ingest_wac_2008/2024.findings-emnlp.713/",
    doi = "10.18653/v1/2024.findings-emnlp.713",
    pages = "12232--12252",
    abstract = "Recent advancements in instruction-tuning datasets have predominantly focused on specific tasks like mathematical or logical reasoning. There has been a notable gap in data designed for aligning language models to maintain topic relevance in conversations - a critical aspect for deploying chatbots to production. We introduce the CantTalkAboutThis dataset to help language models remain focused on the subject at hand during task-oriented interactions. It consists of synthetic dialogues on a wide range of conversation topics from different domains. These dialogues are interspersed with distractor turns that intentionally divert the chatbot from the predefined topic. Fine-tuning language models on this dataset helps make them resilient to deviating from the assigned role and improves their ability to maintain topical coherence compared to general-purpose instruction-tuned LLMs like gpt-4-turbo and Mixtral-Instruct. Additionally, preliminary observations suggest that training models on this dataset also enhance their performance on fine-grained instruction following tasks, including safety alignment."
}