Browsing: LLMs

Amazon Web Services has released an open-source sandbox designed to secure autonomous AI agents through kernel-level isolation and behavioral policy enforcement.

The platform adds support for Meta’s sparse-architecture model, offering granular pricing and flexible deployment options for high-context workloads.

The platform introduces specialized pricing and deployment options for Meta’s latest large-scale model, emphasizing sparse activation and long-context capabilities.

New research from the Stanford University Human-Centered AI Institute reveals that large language models apply inconsistent standards when critiquing reasoning.

The startup introduces Snap, a specialized decision model utilizing sparse attention to process 384,000 tokens for agentic workflows.

Experimental data indicates that deterministic feedback loops can bridge the performance gap between cost-optimized models and frontier systems.

New automated development tools within the ROCm stack reduce manual tuning requirements for high-performance AI and HPC infrastructure.

The new DART platform automates security testing for autonomous agents by deploying adversarial models that learn and adapt to target defenses in real-time.

The new Diamond Adaptive Red Teaming framework provides an automated, multi-turn approach to identifying security vulnerabilities in autonomous enterprise AI systems.

The new flagship model utilizes a sparse activation architecture to balance computational efficiency with high-capacity reasoning across 160 languages.