Related papers: ProbeLogits: Kernel-Level LLM Inference Primitives…

Large language model (LLM) agents on multi-step tasks suffer reasoning degradation, looping, drift, stuck states, at rates up to 30% on hard tasks. Current solutions include hard step limits (abrupt) or LLM-as-judge monitoring (10-15%…

LLM agents with tool access can discover and exploit security vulnerabilities. This is known. What is not known is which features of a system prompt trigger this behaviour, and which do not. We present a systematic taxonomy based on…

LLM debugging agents that consume cloud logs and execute remediation commands are vulnerable to indirect prompt injection through log content. We present LogJack, a benchmark of 42 payloads across 5 cloud log categories, and evaluate 8…

Large language models (LLMs) have shown to be increasingly capable of performing reasoning tasks, but their ability to make sequential decisions under uncertainty only using natural language remains underexplored. We introduce a novel…

We introduce a proxy-analyzer framework for detecting hallucinations in large language models. Instead of looking inside the generating model, our system reads already-generated text through a small locally hosted open-weight model and…

Large language models (LLMs) have achieved widespread adoption across numerous applications. However, many LLMs are vulnerable to malicious attacks even after safety alignment. These attacks typically bypass LLMs' safety guardrails by…

Language models can distinguish between testing and deployment phases -- a capability known as evaluation awareness. This has significant safety and policy implications, potentially undermining the reliability of evaluations that are…

Large Language Models (LLMs) have transformed task automation and content generation across various domains while incorporating safety filters to prevent misuse. We introduce a novel jailbreaking framework that employs distributed prompt…

Tool-using LLM agents must act on untrusted webpages, emails, files, and API outputs while issuing privileged tool calls. Existing defenses often mediate trust at the granularity of an entire tool invocation, forcing a brittle choice in…

AI agents interact with external environments through tool calls, exposing them to attacks like indirect prompt injection that can trigger unauthorized actions. Securing these agents is challenging: they behave autonomously and…

Language models frequently produce plausible yet incorrect reasoning traces that are difficult to verify. We investigate fine-tuning models to use Prolog as an external symbolic reasoning tool, training Qwen2.5-3B-Instruct with Group…

Current research on Vision-Language-Action (VLA) models predominantly focuses on enhancing generalization through established reasoning techniques. While effective, these improvements invariably increase computational complexity and…

The transition of Large Language Models (LLMs) from passive code generators to autonomous agents introduces significant safety risks, specifically regarding destructive commands and inconsistent system states. Existing commercial solutions…

Previous learning-based vulnerability detection methods relied on either medium-sized pre-trained models or smaller neural networks from scratch. Recent advancements in Large Pre-Trained Language Models (LLMs) have showcased remarkable…

Jailbreaks have been a central focus of research regarding the safety and reliability of large language models (LLMs), yet the mechanisms underlying these attacks remain poorly understood. While previous studies have predominantly relied on…

Tool-augmented LLM agents introduce security risks that extend beyond user-input filtering, including indirect prompt injection through fetched content, unsafe tool execution, credential leakage, and tampering with local control files. We…

Designing high-performance kernels requires expert-level tuning and a deep understanding of hardware characteristics. Recent advances in large language models (LLMs) have enabled automated kernel generation, yet most existing systems rely…

We show that LoRA adapters, the dominant distribution format for fine-tuned LLMs, can be reliably backdoored through training data poisoning while preserving baseline task performance. On a Qwen 2.5 1.5B prompt-injection classifier, a small…

Verifiers--functions assigning rewards to agent behavior--have been key to AI progress in math, code, and games. However, extending gains to domains without clear-cut success criteria remains a challenge: while humans can recognize desired…

Large language models have gained widespread prominence, yet their vulnerability to prompt injection and other adversarial attacks remains a critical concern. This paper argues for a security-by-design AI paradigm that proactively mitigates…