CalVerT：透過校準驗證器遙測增強智能體，提升知識密集型任務中的行動與學習

摘要

在知识密集型问答任务中，基于大语言模型的智能体在执行检索和推理行动时，往往无法完整掌握当前答案是否不确定、缺乏支撑或已趋完善，由此引发两种典型失败模式：一是对缺乏依据的答案过度自信，从而降低准确性；二是在已有足够证据的情况下仍然过度检索，造成计算资源浪费。为使智能体更全面地把握其运行状态空间，我们提出**校准验证器遥测（CalVerT）**，通过向智能体状态中注入两类额外遥测数据——校准后的自置信度评分与依据验证器评分——来增强其状态感知能力。实验表明，CalVerT 能在无需训练与基于训练两种场景下提升智能体性能。在四个问答基准测试中，CalVerT 可触发针对智能体过度依赖参数化知识的场景进行检索，从而提升 F1 分数；同时，在智能体已具备足够上下文回答问题的情况下，又能减少冗余检索。我们证明，CalVerT 可在无需训练的前提下增强现有问答框架。此外，CalVerT 同样能改善已训练的系统：仅通过向智能体状态中添加遥测数据，与未添加 CalVerT 遥测但训练方式完全相同的智能体相比，经过强化学习后其性能可观察到显著提升。

English

LLM agents in knowledge intensive question answering take retrieval and reasoning actions with incomplete knowledge about whether their current answer is uncertain, unsupported, or already complete. This produces two failure modes: committing to confident but unsupported answers, which hurts accuracy, and over-retrieving when the evidence in hand already suffices, resulting in wasted compute. To give agents a more complete picture of the state space they are operating in, we introduce calibrated verifier telemetry (CalVerT), which augments the agent's state with additional telemetry: a calibrated self-confidence score and a grounding verifier score. We show that CalVerT can improve agents in both training-free and training-based settings. On four QA benchmarks, we find that CalVerT raises F1 by triggering retrieval in cases where agents over-rely on parametric knowledge, while cutting redundant retrieval in cases where agents have sufficient context to answer. We show that CalVerT can augment existing QA frameworks without training. Moreover, CalVerT also improves trained systems: by simply augmenting an agent's state with telemetry, we observe improvements after reinforcement learning, as compared to an agent with identical training but no CalVerT telemetry.