MedNLP-CHAT Corpus

MedNLP-CHAT Corpus

Update: June 16, 2025, Download file: zip

This page provides access to the corpus used in the NTCIR-18 MedNLP-CHAT Shared Task. The task focuses on evaluating the reliability and safety of chatbot responses to medical questions. The dataset consists of pairs of medical questions and corresponding chatbot responses, each annotated with potential medical, legal, or ethical risks associated with the response. For more details, please refer to the task overview paper and task web page.

References

• Eiji Aramaki, Shoko Wakamiya, Shuntaro Yada, Shohei Hisada, Tomohiro Nishiyama, Lenard Paulo Tamayo, Jingnan Xiao, Axalia Levenchaud, Pierre Zweigenbaum, Christoph Otto, Jerycho Pasniczek, Philippe Thomas, Nathan Pohl, Wiebke Duettmann, Lisa Raithel, and Roland Roller, NTCIR-18 MedNLP-CHAT Determining Medical, Ethical and Legal Risks in Patient-Doctor Conversations: Task Overview, In Proceedings of the 18th NTCIR Conference on Evaluation of Information Access Technologies, pp. 175-189, 2025. DOI: https://doi.org/10.20736/0002002051
• NTCIR-18 MedNLP-CHAT task website: https://sociocom.naist.jp/mednlp-chat/

Licence

CC BY 4.0