Visual-Adversarial-Examples-Jailbreak-Large-Language-Models
Visual-Adversarial-Examples-Jailbreak-Large-Language-Models copied to clipboard
Unispac
Repository for the Paper (AAAI 2024, Oral) --- Visual Adversarial Examples Jailbreak Large Language Models
Hi can you update the command instruction "python minigpt_visual_attack.py --cfg-path eval_configs/minigpt4_eval.yaml --gpu-id 0 --n_iters 5000 --constrained --eps 16 --alpha 1 --save_dir visual_constrained_eps_16" to "python minigpt_visual_attack.py --cfg_path eval_configs/minigpt4_eval.yaml --gpu-_id 0 --n_iters...
hello, i find there are two ways of generating texts in your code: in minigpt_test_manual_prompts_visual_llm.py, you use Chat(), ask() and answer() to generate texts; while in minigpt_inference.py, you use Generator(),...
hello i found that in your code you save images as '.bmp'. i changed the code to save images as '.jpg' and found minigpt4 said the saved adversarial images are...
Hello. I'm interested in your work and have a question about your implementation. When implementing PGD, I noticed that your code uses only the gradient sign, not the gradient value,...
``` Hello~Congratulations on the great work! The configuration is as follows（https://huggingface.co/lmsys/vicuna-7b-delta-v0）： # Vicuna llama_model: "/home/Visual/vicuna-7b-delta-v0-weights" ``` >>> batch_size: 8 0%| | 0/5001 [00:00
Hi! I cloned the vicuna v0 7b repo given in minigpt4 repository and updated the line 16 in minigpt4.yaml file, then I downloaded the minigpt4(vicuna 7b) and updated the path...
Hello! Thanks for your awesome work! I notice that, in the training process, the model is finetuned to directly predict the 66 unsafe sentences without corresponding questions/instructions. However, in the...
Thanks for case. Have you updated the model parameter in the training process?
AssertionError: Model 'minigpt4' has not been registered.