Papers

arXiv 论文提出 recourse validity:LLM 解释被拒请求时常给出正确但无用的理由

Right Reason, Wrong Advice: Evaluating Recourse in LLM Explanations of Denied Requests

Why we picked it

这篇论文戳穿了一个容易被忽略的问题:LLM 解释拒批时给出的理由可能是对的,但改了也照样被拒。作者测了 20 个模型,发现常规“理由正确性”检查基本没用,得让强模型评估“照做能否成功”才行。做审批、风控、权限系统的开发者值得花时间读。

一篇 arXiv 论文提出 recourse validity 概念:当规则系统基于多条独立规则拒绝请求时,LLM 给出的解释即使引用了真实原因,用户照着修改后仍可能被拒。论文用 hitting-set 论证从策略本身预测正确理由何时失效,并在 7 个来自访问控制和美国福利资格判定的规则系统中做了测试,其中 6 个显示单条正确理由几乎不带来 recourse。

Body

实验覆盖 20 个 LLM,常用检查方法(判断引用理由是否正确)在识别这一失败上接近随机水平,与理论预测一致。而让一个较强模型直接评估“照此解释行动能否成功”则能识别问题。更强的推理模型表现更好,但在有 4 条以上独立规则的拒绝场景中仍大量失败;要求列出全部规则对所有开源模型均有帮助。

作者据此建议:资格判定类解释应直接从决策函数构建,并以“行动后是否有效”作为评判标准。这一结论对使用 LLM 生成拒批解释的金融、政务、权限审批系统有直接参考意义。

Original · arXiv cs.AI

Right Reason, Wrong Advice: Evaluating Recourse in LLM Explanations of Denied Requests

Language models increasingly write the explanations people receive when a rule system denies their request. If several rules deny the request independently, an explanation can name a true reason and still leave the person denied after they fix it. We call the missing property recourse validity: changing what the explanation cites, without changing what was asked for, can flip the decision. We compute it exactly, and a hitting-set argument predicts from the policy alone when a correct reason fails. In six of seven rule systems ported from access control and US benefit eligibility, a single correct reason almost never gives recourse when several reasons apply. Across 20 LLMs, checks of whether the cited reasons are correct, computed or asked of an LLM judge, are close to chance at spotting this failure, as the argument predicts. A strong judge asked whether acting on the explanation would work does spot it. Stronger and reasoning models do better but still fail on many denials with four or more independent reasons. Asking for every rule helps every open model with room to improve. Explanations of eligibility decisions should be built from the decision function and judged by whether acting on them works.

Industry Coordinates & Source Verification

This report is categorized under AITOP's "Papers" desk, focusing on US frontier AI labs and Silicon Valley technology leaders. Within our cross-regional analytical framework, it directly maps to AI developer toolchains, coding agents, and engineering workflows. Key entities and technical topics under continuous tracking include LLM, interpretability, arXiv, AI safety.

Source attribution: curated from arXiv cs.AI (@Arunkumar V).