ModelsOfficial

Anthropic 推出 Contextual Retrieval 提升 RAG 准确性

Introducing Contextual Retrieval Sep 19, 2024

Why we picked it

做 RAG 应用的开发者终于有了解决上下文丢失问题的实用方案,错误率直降 67%,建议直接看实现指南。

Anthropic 发布 Contextual Retrieval 技术,通过为每个文本块添加上下文说明,显著提升检索增强生成(RAG)的准确性。

Image · Anthropic: Engineering
Body

传统 RAG 中,孤立文本块常因缺乏上下文导致检索错误,而 Contextual Retrieval 利用 Claude 模型为每个块生成简短描述,使检索更精准。该方法结合 BM25 和嵌入搜索,在多个基准测试中错误率降低 67%。

Anthropic 还提供了高效实现指南,包括使用 prompt caching 降低 1.6% 的成本。

Industry Coordinates & Source Verification

This report is categorized under AITOP's "Models" desk, focusing on US frontier AI labs and Silicon Valley technology leaders. Within our cross-regional analytical framework, it directly maps to US-China foundation model benchmarks and commercialization comparison. Key entities and technical topics under continuous tracking include RAG, Contextual Retrieval, Anthropic, Claude.

Primary source verification: published directly by Anthropic: Engineering with first-party authority. Cross-source corroboration: 11 independent sources have covered this development (including Ethan Mollick, Claude: Blog, Dario Amodei Blog), confirming industry-wide signal strength.