Goodfire says its new ‘inside-out’ monitors catch rogue AI agents at a fraction of the cost

Goodfire推出读取模型内部信号的AI智能体安全监测器，已向Baseten客户开放，称其成本低于由另一模型逐步检查输出的方案。小型探针逐步检测风险，仅在触发警报时交由另一AI模型复查，客户可选择监测风险及记录、人工审核或拒绝请求等响应。