可以在代理工作流程中运行的 LLM 决策组件通常会生成与操作相关的建议或判断以及解释。操作员可以使用指定的因素来监视系统,、诊断错误, 或决定何时升级输出。这种使用假设解释与组件的可观察的决策行为一致。我们测试了对命名因子: 必要性, 的两种解释，这意味着更改因子将改变输出,，而充分性, 则意味着保留它，同时删除其他可更改信息将保留输出。我们在两个综合用例: 中评估这些解释，向客户推荐顾问并判断提示的危害性或风险。模型返回输出以及对其影响最大的三个因素。受控黑盒干预通过测量改变每个因素改变输出,的频率来估计每个因素的必要性分数，并通过测量保留它保留输出的频率来估计充分性分数。在 Claude, GPT, 和 Gemini 系列, 的八个模型中，引用排名与必要性和充分性分数之间的平均斯皮尔曼相关性对于顾问推荐, 为 0.349 和 0.354，对于即时监控为 0.431 和 0.580。此外, 在必要性顾问回复中，有 57.6% 的未引用因素得分高于最低得分引用因素，在充足性情况下，占 58.1%; 相应的提示监控率为 25.8% 和 8.9%。引用的前三个因素包含有用的信息，但不能可靠地识别在必要性或充分性下具有最强测量影响的三个因素。该框架为代理人监督中使用的解释提供了黑盒可靠性检查，同时保留了个人法学硕士决策的范围。

LLM decision components that can operate within agent workflows often produce action-relevant recommendations or judgements together with explanations. Operators may use the named factors to monitor a system, diagnose errors, or decide when to escalate an output. Such use assumes that the explanations agree with the component的 observable decision behaviour. We test two interpretations of the named factors: necessity, meaning that changing a factor would change the output, and sufficiency, meaning that retaining it while removing other changeable information would preserve the output. We evaluate these interpretations in two synthetic use cases: recommending advisors to clients and judging prompts for harmfulness or risk. Models return an output and the top three factors that most influenced it. Controlled black-box interventions estimate a necessity score for each factor by measuring how often changing it changes the output, and a sufficiency score by measuring how often retaining it preserves the output. Across eight models from the Claude, GPT, and Gemini families, the mean Spearman correlations between the cited ranking and the necessity and sufficiency scores are 0.349 and 0.354 for advisor recommendation, and 0.431 and 0.580 for prompt monitoring. Furthermore, an uncited factor scores above the lowest-scoring cited factor in 57.6% of advisor responses under necessity and 58.1% under sufficiency; the corresponding prompt-monitoring rates are 25.8% and 8.9%. The cited top three contain useful information but do not reliably identify the three factors with the strongest measured influence under necessity or sufficiency. The framework provides a black-box reliability check for explanations used in agent oversight while remaining scoped to individual LLM decisions.

科目:人工智能(cs.AI)

Subjects: Artificial Intelligence (cs.AI)