| dataset_info: | |
| features: | |
| - name: answer | |
| dtype: int64 | |
| - name: choices | |
| sequence: string | |
| - name: explanation | |
| dtype: string | |
| - name: question | |
| dtype: string | |
| splits: | |
| - name: train | |
| num_bytes: 303135 | |
| num_examples: 241 | |
| download_size: 153879 | |
| dataset_size: 303135 | |
| configs: | |
| - config_name: default | |
| data_files: | |
| - split: train | |
| path: data/train-* | |
| ## Dataset Description | |
| The pentesting-eval dataset is a multiple-choice collection designed to enhance AI training and evaluation in cybersecurity, particularly for penetration testing. Utilizing OpenAI's GPT-4 Turbo and structured according to the Multitask-Multimodal-Language-Understanding (MMLU) framework, this dataset features a diverse range of questions that mimic various network environments, attack strategies, and real-world penetration testing scenarios. |