摘要：高风险领域中的 AI 生成回答通常表达流畅，却难以验证，尤其是在其包含多步推理而非单一最终答案时。我们提出了一种基于推理、无需参考答案的框架，用于审计 LLM 生成的输出。该方法将生成的推理轨迹分解为若干片段，使用自然语言推理（NLI）标注局部的前提—目标关系，并将这些关系组织成一个超图。随后，确定性的反向 AND-OR 搜索会为各片段分配审计标签，表明每个片段如何在生成回答内部获得支撑。我们在两种场景中评估该框架：使用 Hard2Verify 的演绎数学推理，以及开放式医学推理；后者采用 UroReason——一个新的、由医生标注的基准，包含来自真实临床病例的 LLM 推理轨迹。在这些场景中，我们的 NLI-超图审计相比直接的 LLM-as-judge 基线，提供了更可靠的无参考评估信号。在临床场景中，最先进的 LLM 评审器往往无法识别存在问题的推理片段，会过度接受表达流畅但论证基础薄弱的回答。我们的结果表明，问答评估应考察推断关系如何在整个推理轨迹中组合，而不应仅依赖最终答案或将 LLM 作为验证器。UroReason 将通过 API 提供，我们的代码也将以开源形式发布。

Abstract:AI-generated answers in high-stakes domains are often fluent but difficult to verify, especially when they contain multi-step reasoning rather than a single final answer. We propose a reasoning-based, reference-free framework for auditing LLM-generated outputs. The method decomposes a generated reasoning trace into segments, labels local premise-target relations using Natural Language Inference (NLI), and organizes these relations into a hypergraph. A deterministic backward AND-OR search then assigns segment-level audit labels that indicate how each segment is grounded within the generated response. We evaluate the framework in two settings: deductive mathematical reasoning with Hard2Verify, and open-ended medical reasoning with UroReason, a new physician-annotated benchmark of LLM reasoning traces from real clinical cases. Across these settings, our NLI-hypergraph audit provides a more reliable reference-free evaluation signal than direct LLM-as-judge baselines. In the clinical setting, state-of-the-art LLM judges often fail to identify problematic reasoning segments, over-accepting fluent but weakly grounded responses. Our results show that QA evaluation should account for how inferential relations compose across a reasoning trace, rather than relying only on final answers or LLMs as verifiers. UroReason will be made available through an API, and our code will be released as open source.