arXiv subjects

Dmitrii Volkov

Publications and source records attributed to Dmitrii Volkov.

10 recordsLinked to original sources

arXiv2026

Language Models Can Autonomously Hack and Self-Replicate

We demonstrate that language models can autonomously replicate their weights and harness across a network by exploiting vulnerable hosts. The agent independently finds and exploits a web-application vulnerability, extracts credentials, and deploys an inference server with a copy of its harness and prompt on the compromised host. We test four vulnerability classes: hash bypass, server-side template injection, SQL injection, and broken access control. Qwen3.5-122B-A10B succeeds in 6-19% of attempts, and the smaller Qwen3.6-27B reaches 33% on a single A100. This already matches the current-generation GPT-5.4 and exceeds the prior-generation frontier, where Opus 4 reached 6% and GPT-5 reached 0%. Replicating Qwen weights, frontier models reach 81% (Opus 4.6) and 33% (GPT-5.4). This process chains: a successful replica can repeat it against a new target, producing additional copies autonomously.

cs.CR↗

arXiv2025

GPT-5 at CTFs: Case Studies From Top-Tier Cybersecurity Events

OpenAI and DeepMind's AIs recently got gold at the IMO math olympiad and ICPC programming competition. We show frontier AI is similarly good at hacking by letting GPT-5 compete in elite CTF cybersecurity competitions. In one of this year's hardest events, it outperformed 93% of humans finishing 25th: between the world's #3-ranked team (24th place) and #7-ranked team (26th place). This report walks through our methodology, results, and their implications, and dives deep into 3 problems and solutions we found particularly interesting.

cs.CR↗

arXiv2025

Misalignment Bounty: Crowdsourcing AI Agent Misbehavior

Advanced AI systems sometimes act in ways that differ from human intent. To gather clear, reproducible examples, we ran the Misalignment Bounty: a crowdsourced project that collected cases of agents pursuing unintended or unsafe goals. The bounty received 295 submissions, of which nine were awarded. This report explains the program's motivation and evaluation criteria, and walks through the nine winning submissions step by step.

cs.AI↗

arXiv2025

Demonstrating specification gaming in reasoning models

We demonstrate LLM agent specification gaming by instructing models to win against a chess engine. We find reasoning models like OpenAI o3 and DeepSeek R1 will often hack the benchmark by default, while language models like GPT-4o and Claude 3.5 Sonnet need to be told that normal play won't work to hack. We improve upon prior work like (Hubinger et al., 2024; Meinke et al., 2024; Weij et al., 2024) by using realistic task prompts and avoiding excess nudging. Our results suggest reasoning models may resort to hacking to solve difficult problems, as observed in OpenAI (2024)'s o1 Docker escape during cyber capabilities testing.

cs.AI↗

arXiv2025

Evaluating AI cyber capabilities with crowdsourced elicitation

As AI systems become increasingly capable, understanding their offensive cyber potential is critical for informed governance and responsible deployment. However, it's hard to accurately bound their capabilities, and some prior evaluations dramatically underestimated them. The art of extracting maximum task-specific performance from AIs is called "AI elicitation", and today's safety organizations typically conduct it in-house. In this paper, we explore crowdsourcing elicitation efforts as an alternative to in-house elicitation work. We host open-access AI tracks at two Capture The Flag (CTF) competitions: AI vs. Humans (400 teams) and Cyber Apocalypse (8000 teams). The AI teams achieve outstanding performance at both events, ranking top-5% and top-10% respectively for a total of \$7500 in bounties. This impressive performance suggests that open-market elicitation may offer an effective complement to in-house elicitation. We propose elicitation bounties as a practical mechanism for maintaining timely, cost-effective situational awareness of emerging AI capabilities. Another advantage of open elicitations is the option to collect human performance data at scale. Applying METR's methodology, we found that AI agents can reliably solve cyber challenges requiring one hour or less of effort from a median human CTF participant.

cs.CR↗

arXiv2025

LLM Agent Honeypot: Monitoring AI Hacking Agents in the Wild

Attacks powered by Large Language Model (LLM) agents represent a growing threat to modern cybersecurity. To address this concern, we present LLM Honeypot, a system designed to monitor autonomous AI hacking agents. By augmenting a standard SSH honeypot with prompt injection and time-based analysis techniques, our framework aims to distinguish LLM agents among all attackers. Over a trial deployment of about three months in a public environment, we collected 8,130,731 hacking attempts and 8 potential AI agents. Our work demonstrates the emergence of AI-driven threats and their current level of usage, serving as an early warning of malicious LLM agents in the wild.

cs.CR↗

arXiv2025

Resurrecting saturated LLM benchmarks with adversarial encoding

Recent work showed that small changes in benchmark questions can reduce LLMs' reasoning and recall. We explore two such changes: pairing questions and adding more answer options, on three benchmarks: WMDP-bio, GPQA, and MMLU variants. We find that for more capable models, these predictably reduce performance, essentially heightening the performance ceiling of a benchmark and unsaturating it again. We suggest this approach can resurrect old benchmarks.

cs.LG↗

arXiv2024

BadGPT-4o: stripping safety finetuning from GPT models

We show a version of Qi et al. 2023's simple fine-tuning poisoning technique strips GPT-4o's safety guardrails without degrading the model. The BadGPT attack matches best white-box jailbreaks on HarmBench and StrongREJECT. It suffers no token overhead or performance hits common to jailbreaks, as evaluated on tinyMMLU and open-ended generations. Despite having been known for a year, this attack remains easy to execute.

cs.CR↗

arXiv2024

Hacking CTFs with Plain Agents

We saturate a high-school-level hacking benchmark with plain LLM agent design. Concretely, we obtain 95% performance on InterCode-CTF, a popular offensive security benchmark, using prompting, tool use, and multiple attempts. This beats prior work by Phuong et al. 2024 (29%) and Abramovich et al. 2024 (72%). Our results suggest that current LLMs have surpassed the high school level in offensive cybersecurity. Their hacking capabilities remain underelicited: our ReAct&Plan prompting strategy solves many challenges in 1-2 turns without complex engineering or advanced harnessing.

cs.CR↗

arXiv2024

Badllama 3: removing safety finetuning from Llama 3 in minutes

We show that extensive LLM safety fine-tuning is easily subverted when an attacker has access to model weights. We evaluate three state-of-the-art fine-tuning methods-QLoRA, ReFT, and Ortho-and show how algorithmic advances enable constant jailbreaking performance with cuts in FLOPs and optimisation power. We strip safety fine-tuning from Llama 3 8B in one minute and Llama 3 70B in 30 minutes on a single GPU, and sketch ways to reduce this further.

cs.LG↗