Papers by Judit Dombi
ChatGPT Role-play Dataset: Analysis of User Motives and Model Naturalness (2024.lrec-main)
Copied to clipboard
| Challenge: | Recent advances in interactive large language models like ChatGPT have revolutionized various domains, but their behavior in natural and role-play settings remains underexplored. |
| Approach: | They analyze ChatGPT interactions in a normal way and a role-play setting to examine its behavior in conversational settings. |
| Outcome: | The proposed dataset shows that chatGPT behaves in natural and role-play settings with different user motives and model naturalness. |