In the news
Safety as a Budget: Calibrating the Control Channel of Constrained GRPO for Vision-Language Models
arXiv cs.AI · Published · 3 min read
In 30 seconds
- What happened
- Researchers improved constrained GRPO training for vision-language models by treating safety as explicit budgets on harmful and over-refusal rates rather than indirect reward weights.
- Why it matters
- Matters for engineers building safety-aligned vision-language models who need direct control over acceptable refusal rates and harm thresholds.
- Watch out
- Method requires a calibrated scorer and judge-labeled training data; gains are largest for harms embedded in images, generalization to other domains unclear.
- language model
- reinforcement learning
- grpo
The patterns behind this
Each one covers how the technique works, when it earns its cost, and where it breaks.
The Agent Architect
One pattern, one tradeoff, one production failure story. A short weekly briefing for people building agentic systems.
Weekly email, one-click unsubscribe. We only use your address to send the briefing.