Megadose AI progress, ranked and analyzed.

StepGuard: Learning Step-Level Guardrails with Scalable Supervision and Safety-Utility Balancing

· HF Daily Papers ·
StepGuard checks agent tool actions before execution, cutting attack success by 77.3% in the paper’s agent tests.

The model is trained to judge step-level actions, not just completed agent trajectories. Its StepGen engine creates paired safe and unsafe trajectories with the same context at the risky step. Balance-GRPO is used to reduce both over-blocking and missed unsafe actions. The authors report top average accuracy among open-weight guard models, comparable to GPT-5.4, with utility down 2.8 percentage points. HF Daily Papers' note

score 5

Categories: Research