跳到正文
arXiv cs.CL· Huixue Han, Shangzhi Xu, Anqi Xiong, Yu-Chu Tian·· 9 小时前AI 评分36

聚类越大门槛越低:LLM 增量实体解析中的画像累积效应

Growing Clusters Lower the Bar: Profile Accretion in LLM-Based Incremental Entity Resolution

AI 导读

研究发现,基于 LLM 的增量实体解析中,模型会随聚类增大而降低加入门槛。在 FOPPA 法国采购供应商数据上,将展示的已验证成员从 1 个增至 16 个,六个开源权重模型对相似名称公司的错误匹配率上升数个百分点,而 ROC 曲线下面积仅小幅提升;Claude Sonnet 5 和 Haiku 5.5 上同样出现上升。

正文 · 原文

View PDF HTML (experimental)

Abstract:Incremental entity resolution pipelines ask a large language model (LLM) whether a new record belongs to an existing cluster and show it the records already linked. Analyses of matching oracles assume that the answer depends only on the records compared. For the LLM matchers we test, it depends on cluster size, and the model lowers its bar for joining a cluster as the cluster grows. On French procurement suppliers from FOPPA, with SIREN identifiers as ground truth, we hold the new record fixed and vary only the number of verified members shown. Sixteen members instead of one raise the false match rate on similarly named companies by \dfmrMin--\dfmrMax{} percentage points in six open-weight models, while their area under the ROC curve rises by only \aucGainMin--\aucGainMax. In Claude Sonnet~5 and Haiku~5.5 the rise is \clSonD{} and \clHaiD{} points. The shift persists with street addresses, prompt variants and the BatchER and ComEM formats, and one false merge raises acceptance of the intruder's other records in every open model. It does not arise on Czech and Bulgarian TED notices, where one record with its full address already identifies the company. In online linking on confusable blocks, a single threshold set on labelled data raises recall by a factor of \fxRRmin--\fxRRmax{} for five models with no detectable loss of precision and no extra LLM calls.
Subjects: Computation and Language (cs.CL); Databases (cs.DB)
Cite as: arXiv:2611.22485 [cs.CL]
  (or arXiv:2611.22485v1 [cs.CL] for this version)
  https://doi.org/10.48550/arXiv.2611.22485

arXiv-issued DOI via DataCite (pending registration)

Submission history

From: Huixue Han [view email]
[v1] Tue, 17 Nov 2026 09:42:41 UTC (128 KB)

来源:arXiv cs.CL · arxiv.org