Safety
G-CARB: Graph-Localized Conformal Agent Risk Budget for Compositional Harm
arXiv:2610.05563v1 Announce Type: cross Abstract: Small language model (SLM) agents need safety controls that track consequences across tool calls with little monitoring overhead. A private read, for
arXiv:2610.05563v1 Announce Type: cross Abstract: Small language model (SLM) agents need safety controls that track consequences across tool calls with little monitoring overhead. A private read, for example, becomes a leak when a later action sends that data outside the system. We introduce CARB (Conformal Agent Risk Budget), which calibrates when to stop an agent using a ledger of harm incurred before stopping. Under exchangeable episodes, standard conformal risk control bounds this declared loss in expectation over calibration and a future episode. G-CARB selects scorer evidence along observable dependencies from private sources to outgoing actions. The ledger still covers the entire executed history, and computing the gate score requires no additional language-model inference. On AgentDojo replay with two 14B backbones, G-CARB roughly halves scorer-input records at intermediate risk budgets while improving autonomous task completion relative to full-prefix scoring; random context of the same size achieves similar gains. Controlled examples show how retaining the relevant dependency can further avoid stopping benign work.
Related
- Learning from Failures: Heterogeneous Graph Memory for Small Language Model Tool-Using Agents
- Conformal Policy Control
- IAPO: Input Attribution-Aware Policy Optimization for Tool Use in Small Multimodal Agents
- Calibrate-Then-Delegate: Safety Monitoring with Risk and Budget Guarantees via Model Cascades
- Conformal Risk-Averse Decision Making with Optimized Certainty Equivalent Risk Control
Source: arXiv cs.LG | 2026-10-06