Great work! Thank you for open-sourcing the code.
Would you be willing to share the final JSON files containing the model responses to the 750 harmful instructions from the HADES dataset for LLaVA-1.5 and Gemini-V? I would like to use these results as a baseline for my research.
Thank you very much for your time and help!
Great work! Thank you for open-sourcing the code.
Would you be willing to share the final JSON files containing the model responses to the 750 harmful instructions from the HADES dataset for LLaVA-1.5 and Gemini-V? I would like to use these results as a baseline for my research.
Thank you very much for your time and help!