基于检索的修订与先验注入:探测检索增强的专利权利要求修改
Grounded Revision vs. Prior Injection: Probing Retrieval-Augmented Patent Claim Amendment
- Pukyong National University(釜庆大学)
- Tomocube Inc.(Tomocube 公司)
- Connectionary(Connectionary 公司)
- Teamreboott Inc.(Teamreboott 公司)
机构由 AI 辅助整理,请以论文原文为准。
AI总结:
本研究通过专利权利要求修改场景,提出语料库、探测组和确定性指标,系统检验检索增强生成中检索是否真正支持修订而非注入模板,发现前沿LLM无先验注入行为,检索效应微弱且方向不一致。
AI中文摘要:
检索增强生成广泛应用于专业写作,然而,在“正确”具有可定义含义的场景中,检索是真正为修订提供依据,还是仅仅注入了模板,这一问题很少得到检验。专利权利要求修改恰好提供了这种信号:审查员指出受质疑的权利要求限制并引用在先技术,从而为每个案例提供了基准事实。我们发布了三项成果:(i)一个包含7,385个美国专利商标局审查案例的语料库,其中包含XML对齐的修改前后权利要求、审查意见及引用的在先技术;(ii)一个包含七项探测的测试组,在固定提示框架下比较随机检索与结构匹配检索这两种策略;(iii)一个确定性的五通道评估指标(主指标为C1-C3和C5,补充指标为C4),无需LLM评估。在四个前沿LLM(Claude Sonnet 4、Claude Haiku 4.5、GPT-5.4、GPT-4o-mini)上进行的9,600次预注册调用中,没有测试模型表现出可检测的经典先验注入行为;检索效应较小,且随机检索与结构检索之间的方向不一致;在密集(语义)检索器、检索深度k∈{1,3,5,10}以及基于释义敏感性的接地指标下,零假设保持不变。修订局部性揭示了模板通道无法捕捉的模型特定差异。我们视四单元分类为探索性分析,其中先验注入单元未被占据。
英文摘要:
Retrieval-augmented generation is widely used in professional writing, yet whether retrieval grounds revision or merely injects templates is rarely tested where "correct" has a definable meaning. Patent claim amendment supplies that signal: the examiner names the attacked limitation and cites prior art, providing per-case ground truth. We release three artifacts: (i) a corpus of 7,385 USPTO prosecution cases with XML-aligned pre/post claims, rejection, and cited prior art; (ii) a seven-probe battery comparing random and structural-match retrieval as two policies under a fixed prompt scaffold; (iii) a deterministic five-channel metric (C1-C3 and C5 in main, C4 supplementary) requiring no LLM evaluation. Across 9,600 pre-registered calls on four frontier LLMs (Claude Sonnet 4, Claude Haiku 4.5, GPT-5.4, GPT-4o-mini), no tested model exhibits detectable classical prior-injection behavior; retrieval effects are small and direction-inconsistent between random and structural retrieval, and the null is unchanged under a dense (semantic) retriever, across retrieval depths k in {1,3,5,10}, and under a paraphrase-sensitive grounding metric. Revision locality reveals a model-specific difference that the template channel misses. The four-cell taxonomy, which we treat as exploratory, leaves the prior-injector cell unoccupied.