Unverified50% confidenceFactExact time
The "citation" mechanism in LLM structured extraction requires the model, when outputting each extracted field, to provide the verbatim source snippet in the original document for that value, based on the theoretical rationale that accurately copying source text reduces hallucination risk.
1
Sources
50%
Confidence
Long-term
Relevance
10/19/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedLLM相较传统规则引擎或BI工具的优势在于能处理非结构化文本中的数字引用，如嵌在段落叙述中的百分比或金额73% similarUnverified研究表明，当 LLM 基于结构化输入（如 AST 提取的函数签名和调用关系）而非纯文本推测时，幻觉率会显著降低73% similarUnverifiedThe idea for solving the citation problem is to separate retrieval from generation, letting the model responsible for generation handle only the document content it has already been given, and delegating verbatim quotations to retrieval or string-matching scripts.72% similarUnverifiedSourceCheck的核心理念是用引用的方式替代拷贝，让模型输出极简的元数据来描述引用了哪个文档的哪个段落，而非大段复述原文71% similarUnverifiedRequiring the model in the prompt to tag a source ID after each sentence, or using structured output to return answer and citations, are two ways to force citations at generation time.71% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1071934API
curl https://kongchang.com/api/v1/knowledge/claims/1071934MCP
get_claim(id=1071934)