Large language models (LLMs) like ChatGPT and Gemini have significantly advanced natural language processing, enabling various applications such as chatbots and automated content generation. However, these models can be exploited by malicious individuals who craft toxic prompts to elicit harmful or unethical responses. These individuals often employ jailbreaking techniques to bypass safety mechanisms, highlighting the need for robust toxic prompt detection methods. Existing detection techniques, both blackbox and whitebox, face challenges related to the diversity of toxic prompts, scalability, and computational efficiency. In response, we propose ToxicDetector, a lightweight greybox method designed to efficiently detect toxic prompts in LLMs. ToxicDetector leverages LLMs to create toxic concept prompts, uses embedding vectors to form feature vectors, and employs a Multi-Layer Perceptron (MLP) classifier for prompt classification. Our evaluation on various versions of the LLama models, Gemma-2, and multiple datasets demonstrates that ToxicDetector achieves a high accuracy of 96.39% and a low false positive rate of 2.00%, outperforming state-of-the-art methods. Additionally, ToxicDetector's processing time of 0.0780 seconds per prompt makes it highly suitable for real-time applications. ToxicDetector achieves high accuracy, efficiency, and scalability, making it a practical method for toxic prompt detection in LLMs.

Compilers are one of the most fundamental software systems. A large number of software systems rely on compilers for execution. Compiler bugs can significantly hinder software developers from diagnosing issues within their software. Therefore, it is essential to ensure the correctness of compilers and to isolate and fix compiler bugs. Isolating bugs within compilers is challenging due to compilers' complexity and large codebase. The prior studies on compiler bug isolation struggle to generate sufficient test cases for bug isolation and are not effective enough.

GOST, known as GOST-28147-89, was standardized as the Russian encryption standard in 1989. It is a lightweight-friendly cipher and suitable for the resource-constrained environments. However, due to the simplicity of GOST’s key schedule, it encountered reflection attack and fixed point attack. In order to resist such attacks, the designers of GOST proposed a modification of GOST, namely, 2-GOST. This new version changes the order of subkeys in the key schedule and uses concrete S-boxes in round function. But regarding single-key attacks on full-round 2-GOST, Ashur et al. proposed a reflection attack with data of 2 32 on a weak-key class of size 2 224 , as well as the fixed point attack and impossible reflection attack with data of 2 64 for all possible keys. Note that the attacks applicable for all possible keys need the entire plaintext space. In other words, these are codebook attacks. In this paper, we propose single-key attacks on 2-GOST with only about 2 32 data instead of codebook. Firstly, we apply 2-dimensional meet-in-the-middle attack combined with splice-cut technique on full-round 2-GOST. This attack is applicable for all possible keys, and its data complexity reduces from previous 2 64 to 2 32 . Besides that, we apply splice-cut meet-in-the-middle attack on 31-round 2-GOST with only data of 2 32 . In this attack, we only need 8 bytes of memory, which is negligible.