Süha Orhun Mutluergil

Caglar Uysal, Baturay Birinci, Süha Orhun Mutluergil et al.

Large Language Models (LLMs) have rapidly evolved, transforming industries by automating complex tasks and generating human-like content. However, as their adoption accelerates, prompt injection vulnerabilities have become increasingly apparent. Malicious actors exploit these weaknesses to generate phishing emails, deceptive websites, nd malware, posing serious security risks. This paper presents an empirical evaluation of six state-of-the-art LLMs (DeepSeek, GPT, Gemini, Grok, Llama, and Qwen) under diverse adversarial prompt scenarios, including direct and multi-stage obfuscated attacks across multiple languages and character encodings. The proposed framework measures how effectively current LLMs resist manipulation into performing harmful actions. Our findings reveal systematic vulnerabilities across all tested models. Even direct prompt injections frequently induce the generation of phishing content, websites, and malware, while elaborate prompts achieve even higher malicious compliance rates, particularly for phishing. Models such as DeepSeek, Gemini, and Grok show especially high susceptibility under complex instructions. Notably, non-English languages consistently exhibit higher compliance rates than English, exposing significant gaps in multilingual safety alignment. Although simple character encodings reduce malicious outputs, they do not eliminate them. These results highlight persistent challenges in LLM safety and underscore the urgent need for stronger defenses and improved security mechanisms to support the ethical and secure deployment of LLMs in cybersecurity sensitive contexts.

Suha Orhun Mutluergil, Alperen Dogan

Refinement proofs verify an implementation by showing that its behaviours are subsumed by a simpler specification, on which safety properties are easier to establish. We study how such proofs interact with the axiomatic program logics used to verify the specification. We first give a uniform account of Hoare, Incorrectness, Lisbon, and Necessary-Preconditions logic, classified by the direction in which each constrains a transition and by whether it over- or under-approximates its target set. We then show that simulation relations transfer state-based safety properties: a forward simulation carries a Hoare (inductive) invariant of the specification to one of the implementations, and forward and backward simulations both carry ordinary invariants, via the pre-image of the relation. Finally, we characterize, within these logics, when a relation is a simulation, forward simulations by the validity of Hoare or Lisbon triples, backward simulations by Necessary-Preconditions or Incorrectness triples, so that the simulation obligation reduces to a triple in an off-the-shelf functional logic. We illustrate the development with a concurrent counter, transporting a safety bound from an atomic sequential specification to a Left--Right implementation through an intermediate nondeterministic-concurrent counter, with a forward simulation on one side and a backward simulation on the other.

Ege Aktemur, Ege Zorlutuna, Kaan Bilgili et al.

We introduce a new approach in distributed deep learning, utilizing Geoffrey Hinton's Forward-Forward (FF) algorithm to speed up the training of neural networks in distributed computing environments. Unlike traditional methods that rely on forward and backward passes, the FF algorithm employs a dual forward pass strategy, significantly diverging from the conventional backpropagation process. This novel method aligns more closely with the human brain's processing mechanisms, potentially offering a more efficient and biologically plausible approach to neural network training. Our research explores different implementations of the FF algorithm in distributed settings, to explore its capacity for parallelization. While the original FF algorithm focused on its ability to match the performance of the backpropagation algorithm, the parallelism aims to reduce training times and resource consumption, thereby addressing the long training times associated with the training of deep neural networks. Our evaluation shows a 3.75 times speed up on MNIST dataset without compromising accuracy when training a four-layer network with four compute nodes. The integration of the FF algorithm into distributed deep learning represents a significant step forward in the field, potentially revolutionizing the way neural networks are trained in distributed environments.