Large language model accurately predicts online chat derailments

Celebrity Gig
Examples of conversational derailment on Wikipedia Talk Pages: On the left is an example of a discussion that concluded without derailment, while on the right is an example of a derailed conversation. In the case on the right, the final comment ends with an aggressive remark directed at the other person: “…wasting other people’s time is simply rude.” Credit: IEEE Access (2025). DOI: 10.1109/ACCESS.2025.3554548

Online chat rooms and social networking platforms frequently experience harmful behavior as discussions drift from their intended topics toward personal conflict. Traditional predictive models typically depend on platform-specific data, limiting their applicability and increasing implementation costs.

In a new study, researchers at the University of Tsukuba applied a zero-shot prediction method to LLMs to detect conversational derailments. The performance of various untrained LLMs was compared to that of a deep learning model trained on curated datasets. The results showed that untrained LLMs achieved comparable, and in some cases superior, accuracy.

READ ALSO:  How can Europe boost its domestic battery production and what policy actions are needed?

These findings, published in the journal IEEE Access, suggest that platform operators can implement effective moderation tools at reduced cost by leveraging general-purpose LLMs, supporting healthier online communities across diverse platforms.

More information:
Kenya Nonaka et al, Zero-Shot Prediction of Conversational Derailment With Large Language Models, IEEE Access (2025). DOI: 10.1109/ACCESS.2025.3554548

READ ALSO:  Bahamas files emergency motion in FTX case for access to customer data

Provided by
University of Tsukuba


Categories

Share This Article
Leave a comment