Related papers: Committed SAE-Feature Traces for Audited-Session S…

Automated fault localization requires connecting an observed test failure to the responsible method across thousands of candidates--a task that purely statistical approaches handle with limited precision and that LLMs cannot yet handle at…

Do stock safety-aligned language models and their uncensored or abliterated derivatives behave differently when run as autonomous security agents? Single-turn refusal benchmarks cannot answer this question: security agents must inspect…

The deployment of large language models (LLMs) on third-party devices requires new ways to protect model intellectual property. While Trusted Execution Environments (TEEs) offer a promising solution, their performance limits can lead to a…

Patient-voiced clinical-triage benchmarks report high under-triage rates for consumer LLMs for constrained multiple-choice output, yet the same cases score differently with free-text. We ask whether output format changes the model's…

Large Language Models (LLMs), especially their compact efficiency-oriented variants, remain susceptible to jailbreak attacks that can elicit harmful outputs despite extensive alignment efforts. Existing adversarial prompt generation…

Insider threat detection assumes that an adaptive insider leaves behavioral residue distinguishing them from legitimate users. We test this assumption against an LLM-driven adaptive insider in a controlled multi-agent simulator. Our…

An upstream task for software bill-of-materials (SBOMs) is the accurate localization of the patch that fixes a vulnerability. Nevertheless, existing work reveals a significant gap in the CVEs whose patches exist but are not traceable.…

Sparse autoencoders (SAEs) are an unsupervised method for learning a sparse decomposition of a neural network's latent representations into seemingly interpretable features. Despite recent excitement about their potential, research…

Self-supervised landmark estimation is a challenging task that demands the formation of locally distinct feature representations to identify sparse facial landmarks in the absence of annotated data. To tackle this task, existing…

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%…

We characterize a compositional architecture of literary primitives in two instruction-tuned large language models (Llama 3.1 8B-Instruct and Gemma 2 9B-IT) via sparse autoencoders on mid-depth residual streams. Four feature classes emerge:…

Backdoor attacks on language models pose a significant threat to AI safety, where models behave normally on most inputs but exhibit harmful behavior when triggered by specific patterns. Detecting such backdoors through mechanistic…

Vision-language models (VLMs) have advanced rapidly and are increasingly deployed in real-world applications, especially with the rise of agent-based systems. However, their safety has received relatively limited attention. Even the latest…

Biosecurity evaluations of language models typically ask whether models produce hazardous output. This paper asks a complementary question: when a model refuses, is that refusal structurally sound, or does it disappear under modest changes…

We present a novel class of jailbreak adversarial attacks on LLMs, termed Task-in-Prompt (TIP) attacks. Our approach embeds sequence-to-sequence tasks (e.g., cipher decoding, riddles, code execution) into the model's prompt to indirectly…

Gating mechanisms have been widely utilized, from early models like LSTMs and Highway Networks to recent state space models, linear attention, and also softmax attention. Yet, existing literature rarely examines the specific effects of…

Despite their remarkable success, large language models (LLMs) have shown limited ability on safety-critical code tasks such as vulnerability detection. Typically, static analysis (SA) tools, like CodeQL, CodeGuru Security, etc., are used…

License Assisted Access (LAA) LTE (LTE-LAA) is a new type of LTE that aggregates the licensed LTE bands with the unlicensed bands via carrier aggregation. To operate in unlicensed bands, LTE-LAA adopts the listen-before-talk policy and…

We introduce an evaluation framework of 500 C verification tasks across five property types (memory safety, overflow, termination, reachability, data races) built on SV-COMP 2025, and evaluate 14 models across six families. We find that…

A protocol for two-party secure function evaluation (2P-SFE) aims to allow the parties to learn the output of function $f$ of their private inputs, while leaking nothing more. In a sense, such a protocol realizes a trusted oracle that…