Recent explorations with commercial Large Language Models (LLMs) have shown that non-expert users can jailbreak LLMs by simply manipulating their prompts; resulting in degenerate output behavior, privacy and security breaches, offensive outputs, and violations of content regulator policies. Limited studies have been conducted to formalize and analyze these attacks and their mitigations. We bridge this gap by proposing a formalism and a taxonomy of known (and possible) jailbreaks. We survey existing jailbreak methods and their effectiveness on open-source and commercial LLMs (such as GPT-based models, OPT, BLOOM, and FLAN-T5-XXL). We further discuss the challenges of jailbreak detection in terms of their effectiveness against known attacks. For further analysis, we release a dataset of model outputs across 3700 jailbreak prompts over 4 tasks.

🗂 Paper: https://arxiv.org/pdf/2305.14965

@scopeofai
@LLM_learning

Deep-Live-Cam

Real time face swap and one-click video deepfake with only a single image

Creator: Hacksider
Stars ⭐️: 50,498
Forked by: 7,491

Github Repo:
https://github.com/hacksider/Deep-Live-Cam

@LLM_learning

With so many LLM papers being published, it's hard to keep up and compare results. This study introduces a semi-automated method that uses LLMs to extract and organize experimental results from arXiv papers into a structured dataset called LLMEvalDB. This process cuts manual effort by over 93%. It reproduces key findings from earlier studies and even uncovers new insights—like how in-context examples help with coding and multimodal tasks, but not so much with math reasoning. The dataset updates automatically, making it easier to track LLM performance over time and analyze trends.

📂 Paper: https://arxiv.org/pdf/2502.18791

▫️@scopeofai
▫️@LLM_learning