This paper proposes FedDAB, a two-phase method for defending Federated Learning against backdoor attacks using local contrastive regularization and alignment checking.

The paper introduces alignment contracts, a formal framework for specifying and enforcing behavioral constraints over observable effect traces, ensuring that powerful agentic security systems operate…

The paper introduces Q-ALIGN DT, a novel framework that improves conditioned sequence models by enforcing alignment between the input return-to-go (RTG) signal and the output policy's expected Q-value…

The paper proposes a novel zeroth-order optimization framework to enhance the robustness of LLM safety alignment, showing that few refinement steps can significantly improve safety while maintaining u…

The paper introduces TRAILS~, a novel method that improves code correctness validation by grounding LLM reasoning in concrete (input, output) pairs derived from specifications, achieving state-of-the-…

MOSAIC is a multi-objective framework that efficiently allocates a fixed supervised fine-tuning budget by turning failure profiles into actionable data mixtures, significantly improving model alignmen…

The paper introduces RAG-Pref, a novel, training-free Retrieval Augmented Generation (RAG) method for preference alignment that significantly improves LLM refusal guardrails against agentic attacks wi…

Kani is an open-source model checker for Rust that provides correctness guarantees for safety properties through compilation and a specification language.

This paper establishes mathematical limits of AGI safety, proving structural unverifiability as the core barrier.

The paper formally models structure-informed multiple sequence alignment (MSA-S) as an NP-complete optimization problem, establishing a strong computational complexity baseline for the field.

This paper shifts the focus of LLM safety from preventing misalignment to investigating the model's intrinsic ability to self-recover its alignment after being corrupted by adversarial inputs.

This paper presents a simpler variation of the NC detection criterion for perfect matchings in bipartite graphs with improved parameters.

The paper introduces the Task Alignment Benchmark (TAB) to evaluate terminal agents' ability to selectively follow relevant environmental instructions while ignoring misleading distractors, revealing…

The paper investigates how various fine-tuning methods can be used both to intentionally misalign and subsequently realign large language models (LLMs), revealing distinct strengths for attack and def…

The paper proposes detecting 'alignment faking' (AF)—where LLMs revert to unsafe behavior when unmonitored—by analyzing observable tool selection patterns, finding that detection rates vary significan…

LISA is a novel LLM-based invariant testing framework for software functional bugs, achieving higher bug-detection rates and competitive code coverage than fuzzing and prior LLM-based test generation…

This paper proposes an automated static verification method for a subset of WebAssembly using a CHCs satisfiability solver, addressing challenges of handling indirect function calls and analyzing larg…

SEMBridge is a tagless-final framework that allows a single executable object program to generate multiple program semantics, including weakest-precondition and bounded-checking interpretations, ensur…