| # Oracle-RL | |
| Oracle-RL is a target model developed to evaluate training data detection methods in realistic, large-scale reinforcement learning (RL) settings. | |
| The corresponding benchmark dataset is [WildTrace-RL](https://huggingface.co/datasets/SUSTech/WildTrace-RL). | |
| **Project Homepage:** [WildTrace](https://jaydencoolcc.github.io/wildtrace-website/) |