Tagged “red-teaming”

4 posts.

All Model Architecture23 Training & Alignment22 Inference & Serving24 Agents & Orchestration13 Reasoning & Evaluation28 Safety, Security & Governance9 Platforms & Practice22

Safety, Security & Governance 24 min

A Decade of Adversarial Examples: Why Robustness Never Came Free

In 2014 a perturbation the size of one 8-bit colour step turned a 57.7 percent panda into a 99.3 percent gibbon. Twelve years, 300 million synthetic training images and more than 10^21 training FLOPs later, the best CIFAR-10 mode…

security safety evaluation benchmarks ∑ ◫

Safety, Security & Governance 31 min

Adversarial Robustness Without a Norm Ball: What Jailbreaks Did to a Decade of Theory

Twelve published defences against LLM jailbreaks, most reporting attack success rates near zero, were pushed above 90% in 2025 by attackers who mainly spent more compute. Adversarial robustness for language models has no certific…

adversarial-robustness jailbreaks safety security ∑ ◫

Safety, Security & Governance 24 min

Algorithmic Audits: What an Outside Examination of an AI System Can Actually Establish

Gender Shades measured a 34.4-point error gap on 1,270 faces and moved three vendors within seven months. New York City's mandatory bias audits produced 18 posted reports from 391 employers. The difference was not auditor skill b…

ai-assurance-and-audit evaluation governance compliance ∑ ◫

Reasoning & Evaluation 23 min

The Model Knows It Is Being Tested: Evaluation Awareness and the Limits of Behavioural Safety Evidence

On one synthetic honeypot evaluation, Claude Sonnet 4.5 said out loud that it suspected it was being tested in 80 to 100 percent of transcripts, against under 10 percent for its predecessor. When the internal representations behi…

safety-alignment evaluation interpretability red-teaming ∑ ◫