overfeed.news
Anuncio

NexTalk

Automatiza la primera respuesta. Mantén el toque humano.

Chat en vivo con agentes de IA en tu modelo y tu clave de API, con traspaso a tu equipo.

Empezar gratis

Tema

Seguridad y regulación

2482documentos

209últimos 7 días

arXiv AI Papers

Online Inference in Distributional Temporal-Difference Learning

We study online statistical inference for functionals of the return distribution under a fixed policy. The return distribution is estimated by nonparametric distributional temporal-difference learning from a single Markov trajectory. For the Polyak--Ruppert averaged estimator, we prove that its root-T error converges weakly to a centered Gaussian random element in Cramér space. We also prove that, conditionally on the observed trajectory, the root-T difference between the bootstrap and original averages converges weakly to the same Gaussian limit. These results justify bootstrap inference for smooth statistical functionals, including variance, CVaR, expected shortfall, and expectiles. For nonsmooth statistical functionals, we develop a local asymptotic theory for the estimated return CDF over T^{-1/2}-neighborhoods of finitely many thresholds, together with its bootstrap analogue. This theory allows us to conduct inference for nonsmooth statistical functionals characterized by CDF equations, including return quantiles.

en
arXiv AI Papers

Tripwire: Triggering Aligned Refusal via Statistically Certified Safety Neurons

Neuron- and path-level interventions offer the finest-grained route to defending large language models (LLMs) against jailbreak attacks, yet existing methods fall short of this promise, i.e., they often compromise model utility significantly. Specifically, one line of work suppresses toxic neurons to erase harmful semantics, but since such semantics are distributed across the network, blocking every pathway forces a large intervention footprint. An alternative line of research focus on identify safety neurons using external classifiers. While promising, the existing approaches suffer from compromising neurons that are important for the model utility as well. Moreover, both approaches remain always on and thus perturb every benign request even when no attack is present. To address these limitations, we present {}, a training-free defense that first identifies safety-specific neurons through per-neuron hypothesis tests under false-discovery-rate control together with a utility-specificity filter. Based on this identification, a trigger-style clamp holds the selected neurons at their harmful-conditional mean activations, injecting an internal harmful-input signal that triggers the refusal behavior learned during alignment. The clamp is then realized by two provably equivalent deployment modes, namely a detector-gated inference-time intervention and an offline bias-patch weight edit. Extensive experiments across four safety-aligned LLMs and four representative attacks demonstrate that {} reduces the average attack success rate to at most 2.0\% while incurring a utility drop of only 0.5\% to 5.3\% on MT-Bench, the smallest among all defenses. Code is available at https://anonymous.4open.science/r/Tripwire-65C4.

en
arXiv AI Papers

AgentRewind: Recoverable Execution for Long-Horizon LLM Agents

Many real-world tasks require LLM agents to interact with their environments over long execution horizons. Errors that occur early in execution may propagate through both the agent context and environment state, and their effects may be difficult to reverse through subsequent actions. Existing methods mainly seek to reduce such errors through plan refinement and safety checks but provide little support after errors occur. To enable recovery during long-horizon execution, we present AgentRewind, a runtime recovery framework that records aligned checkpoints of the agent context and controlled environment, allowing agents to return to an earlier state and resume execution with information from previous attempts. We also construct MettleBench, a benchmark for evaluating task completion and partial progress on long-horizon engineering assignments containing a series of related requirements. Experiments across tasks, multiple models, execution strategies, and agent harnesses show that AgentRewind improves task success rate and average checklist progress over the compared baselines.

en
AWS Machine Learning Blog

Pay with confidence: How Solv Labs built verifiable, auditable agent payments on Amazon Bedrock AgentCore payments

Solv Labs built a governed agent-payments workflow on Amazon Bedrock AgentCore payments, where every transaction is authorized, attested in an AWS Nitro Enclave, priced for risk, and anchored to a public blockchain before settlement. See how the pattern gives enterprises a verifiable, auditable trail for autonomous agent payments in regulated environments.

en
NVIDIA AI Blog

AI Leaders Propose SAFE Guidelines for Cybersecurity Transparency

Members of the Open Secure AI Alliance — now more than 120 organizations strong — are developing new guidelines to strengthen agentic AI cybersecurity as the annual Black Hat conference begins in Las Vegas today. The Linux Foundation today shared a Request for Comments on Shared AI Findings Exchange (SAFE), a proposed set of guidelines […]

en