Topic · 10 recaps

Alignment & Safety

Research on getting models to do what we actually want — honest, harmless, and helpful behavior — covering RLHF, constitutional methods, adversarial robustness, and broader safety theory.

Posts Questions

Get Started

Topic · 10 recaps

Alignment & Safety

Research on getting models to do what we actually want — honest, harmless, and helpful behavior — covering RLHF, constitutional methods, adversarial robustness, and broader safety theory.

Posts Questions

Sort

Newest

SchemeArena: Factorized Stress Testing of Scheming in LLM Agents

Agents · Sep 8

Steering Geometry: Validating Human Value Geometry in LLM Steering Space

Alignment · Sep 5

Your Model's Chain of Thought Is a Sensor, Not a Security Boundary

Alignment · Sep 4 · 14:56

Knowing What Not to Answer: Selective Non-Compliance in Vision-Language Models

Alignment · Sep 4

Refuse without Refusal: A Structural Analysis of Safety-Tuning Responses for Reducing False Refusals in Language Models