This paper systematically studies how soft errors propagate during Large Language Model (LLM) inference using a novel fault-injection framework, providing critical insights and mitigation strategies f…

This paper introduces a failure-aware observability framework to diagnose wasted computation in multi-agent LLM systems by mapping recurring failure modes to online trace signals.

The paper introduces 'abliteration,' a weight editing technique that successfully bypasses the refusal mechanism of safety-aligned Code LLMs, enabling scalable synthesis of vulnerable code from safe i…

The paper introduces False Security Confidence (FSC), a new metric to measure the inherent prevalence of security vulnerabilities in code generated by LLMs that are otherwise functionally correct, eve…

The paper addresses the gap in understanding real-world LLM-in-the-loop vulnerabilities by creating the LLMCVE dataset and demonstrating that these vulnerabilities are significantly harder to repair t…

This paper identifies the 'Format-Reliability Gap'—where LLMs know about code vulnerabilities but generate insecure code anyway—and proposes a localized, per-vulnerability steering vector fix that sig…

The paper identifies and measures a critical failure mode where LLM agents violate policies by losing or corrupting directive-bearing state during the process of assembling the decision context, and p…

The paper demonstrates that many instruction-tuned language models suffer from 'silent commitment failure,' meaning they can produce confidently incorrect outputs without any warning signal, and intro…

FALAT is a diagnostic framework that treats failure attribution in complex LLM agent trajectories as a dependency-guided search problem, successfully identifying both the responsible agent and the dec…

This paper proposes a method to recover recoverability structure from failed traces of post-trained language models, enabling test-time routing and post-training analysis.

The paper introduces POIROT, a novel protocol that uses the agents within a multi-agent system itself to diagnose and detect failures, demonstrating superior performance over traditional evaluation me…

The paper conducts an interpretability-driven safety audit of eight state-of-the-art LLMs, demonstrating that while interpretability-based steering is a powerful auditing tool, model robustness varies…

This paper addresses the critical need for trustworthy LLMs in science by proposing a comprehensive, multi-layered defense framework and methodology to evaluate unique scientific vulnerabilities.

The paper benchmarks current frontier computer-using agents against hand-crafted attacks, finding that while they are highly safe in browser tasks, this safety does not generalize to other domains lik…

The paper demonstrates that standard LLM evaluation metrics overestimate performance because they fail to account for the stability of outcomes, showing a significant gap between reported pass rates a…

The paper introduces LACUNA, a novel programming model that allows LLM agents to write code that shapes the runtime environment while maintaining strong type-checking safety guarantees.

VulKey introduces a novel LLM-based framework that uses a hierarchical abstraction of expert security knowledge to guide automatic vulnerability repair, achieving state-of-the-art performance on real-…

This systematic mapping survey reviews label-efficient approaches for code vulnerability detection, synthesizing five paradigm families and providing a decision guide to navigate trade-offs.

The paper introduces Refute-or-Promote, an adversarial multi-agent review system that significantly improves the precision of LLM-assisted defect discovery by filtering out false positives.

The paper introduces SafeRedirect, a system-level defense that prevents frontier LLMs from generating harmful content during legitimate tasks that structurally require it, significantly reducing unsaf…