New Research: LLM Judges Unstable, RAG Privacy, and More
New arXiv papers cover LLM judge instability (Wiggle Framework), privacy-preserving RAG (SEAG), multi-agent pricing, transformer expressiveness, constraint saturation, Dual-Flow Transformers, governed memory, and edge scheduling. Findings: judges flip verdicts 25-91% under pressure; SEAG hides sensitive entities with >74% accuracy; instruction following collapses beyond 5-6 constraints.
Why this appeared
- Multiple sources
These labels describe how the story was ranked for the feed — not editorial endorsements.
Sources and evidence
- arXiv
- arXiv
- arXiv
- arXiv
- arXiv
- arXiv
- arXiv
- arXiv
- arXiv
- arXiv
- arXiv
- arXiv
- arXiv
- arXiv
- arXiv
- arXiv
- arXiv
- arXiv
- arXiv
- arXiv
- arXiv
- arXiv
- arXiv
- arXiv
- arXiv
- arXiv
- arXiv
- arXiv
- arXiv
- arXiv
- arXiv
- arXiv
- microsoft.com
- arXiv
Summary last validated Aug 15, 2026
Reader actions
Report an issue
Use this for an incorrect summary, wrong source, duplicate story, or wrong category. Submissions are private and do not change the story automatically.
Related coverage
- security
Microsoft urges early testing of post-quantum certificate ecosystems
Microsoft published guidance urging organizations to begin testing certificate ecosystems for post-quantum authentication now, highlighting its PQC TLS Pilot Program as a way to advance future readiness. The post appeared on the Microsoft Security Blog.
- research
NIST Study Maps How Toxic Adulterants in Fentanyl Vary Across US
NIST researchers analyzed fentanyl samples nationwide and found that the toxic substances mixed into the drug vary by region and shift over time. The study aims to help first responders and law enforcement identify dangerous adulterants in local supplies, which the agency says can help save lives.
- research
NVIDIA teaches robots to assemble GB300 tester trays
NVIDIA describes how it taught robots to assemble tester trays for its Grace Blackwell GB300 superchip, which powers AI training and inference. The work required skilled physical labor in factories worldwide. NVIDIA says the project offered lessons in robot learning, mechanical intelligence, and traditional engineering.
- agents
Microsoft Research releases Agent Lightning v1.0 agentic RL framework
Microsoft Research released Agent Lightning v1.0, a 3,500-line lightweight agentic reinforcement learning framework. It connects existing AI agents to RL training so they can be improved without rebuilding them, addressing the difficulty of training agents whose tools, context, and decision-making are managed by complex frameworks.
- security
Microsoft FORGE Lab Scales AI Vulnerability Research to Linux Kernel
Microsoft Security's FORGE Lab detailed three lessons from its frontier AI vulnerability research, describing how it is scaling the effort from Windows to the Linux kernel. The work focuses on using AI to find security flaws across a broader range of software targets.