Datasets:
Request access to sharetrawl
This repository is publicly accessible, but you have to accept the conditions to access its files and content.
This dataset holds real conversations that people published from ChatGPT, Claude, Gemini, Grok and other assistants, collected from the open web. It contains personal information people chose to share and material that is offensive, explicit or harmful. Access is approved by hand, for research. By requesting access you agree to: use the dataset under CC BY-NC 4.0 (non-commercial, attribution), the terms the embedded ShareChat rows carry; make no attempt to identify the people in these conversations or to infer sensitive personal data about them; apply your own filtering and safety measures before training or evaluating models on it; respect the terms of service of the platforms the conversations came from; not redistribute the dataset; and delete any conversation the maintainers notify you has been removed.
Log in or Sign Up to review the conditions and access this dataset content.