Alignment

December 27, 2025

Steering LLMs: Real-Time Model Control Without Retraining (Guide, SEO, Practical)

Discover how steering lets you guide an LLM at inference time, without retraining. Concepts, use cases, frameworks, and best practices for robust, indexable content.

Read article

October 26, 2025

Stress Testing AI Alignment: Why Models That Seem Safe Might Just Be Good at Taking Tests

A highly capable AI system might secretly pursue misaligned goals—a phenomenon referred to as 'scheming'. This research explores whether current alignment techniques are truly effective or if models are just getting better at passing evaluations.

Read article