arXiv cs.AI· S M Rayeed, Charles V. Stewart, Yao Ma·· 10 小时前AI 评分31
CLADE：检索增强语言模型在证据粒度不足时的类别级作答现象
CLADE: Category-Level Answering from Deficient Evidence in Retrieval-Augmented Language Models
AI 导读
研究者提出 CLADE 现象：检索增强语言模型读到只提到上位类别的粗粒度段落时，会直接作答而非弃答或继续检索。在 PrimeKG 生物医学问题上，仅含疾病家族的段落使弃答率比无信息段落降低 13.8 个百分点，猜测约一半错误；将精确疾病名替换为家族名比删除该句降低弃答率 17.4 个百分点。
正文 · 原文
Abstract:Retrieval-augmented systems often retrieve passages that are accurate and relevant but state the answer at a coarser level of granularity than the question requires. For example, a question may ask which specific disease a drug treats (familial hyperlipidemia), while the retrieved passage mentions the broader disease family (hyperlipidemia). Prior work on context sufficiency treats such coarse passages simply as insufficient. We argue that they form a distinct case, which we call resolvable insufficiency: text alone cannot settle the fine-grained answer, but the broader category it mentions maps to a short list of candidate answers in a domain ontology. When language models read these coarse, family-only passages, they exhibit a phenomenon we call Clade (Category-Level Answering from Deficient Evidence): they answer instead of abstaining or searching further. On biomedical questions from PrimeKG, family-only passages reduce abstention by 13.8 points compared to uninformative passages, and the resulting guesses are wrong about half the time. A counterfactual test confirms the cause: replacing an exact disease name with its broader family name lowers abstention by 17.4 points compared to deleting the sentence. Simple fixes like confidence filtering, scaling to a larger model, and explicitly instructing the model to abstain cannot resolve Clade, and this replicates on ecological species interactions in GloBI. Finally, we test whether an ontology can turn coarse passages into exact answers. Given the correct family, an ontology list includes the true answer in 98.5% of cases, but string matching finds that family on the page for only 19.8%. Letting the model choose the family raises this to 30.0% and end-to-end recovery from 10.0% to 26.8%. Overall, our results show when language models should stop guessing and consult an ontology, and highlight which steps of that process remain unsolved.
| Subjects: | Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2611.14355 [cs.AI] |
| (or arXiv:2611.14355v1 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2611.14355 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: S M Rayeed [view email]
[v1]
Tue, 10 Nov 2026 17:16:06 UTC (1,688 KB)
来源：arXiv cs.AI · arxiv.org