liveBook · Manning

Anchor vector

As explained in Retrieval Augmented Generation, The Foundational Ideas

An anchor vector is a numerical summary representation formed from important sentences selected from retrieved passages.

An anchor vector is a summary representation built from the most important sentences in retrieved passages. In multi-document question answering, it concentrates fine-grained evidence before the decoder processes the remaining context, improving grounding and reducing decoder memory reads. section 5.6.1

A multi-granularity fusion process forms and uses the vector in four steps:

  • Re-rank the retrieved passages. section 5.6.1
  • Select the most important sentence from each top-ranked passage. section 5.6.1
  • Aggregate those sentences into an anchor vector. section 5.6.1
  • Prune less important passages and give the decoder the vector together with the remaining passages. section 5.6.1

Evidence-focused fusion

Suppose several retrieved passages contain mixed-quality evidence. The system keeps one important sentence from each top-ranked passage, combines those sentences into the anchor vector, removes weaker passages, and lets the decoder read a more focused context. The selected evidence also helps show which passages supported the answer. section 5.6.1

How Retrieval Augmented Generation, The Foundational Ideas covers it

How deep it goes
Explains a four-step formation and fusion procedure, but does not detail the vector's numerical construction.
Examples
Uses reranked passages, selected key sentences, an anchor vector, pruned passages, and decoder context.
What it assumes
Assumes retrieved passages and a decoder-based fusion pipeline, while clarifying how evidence is reduced from passages to sentences.
What you can do after
Evaluate supporting evidence and focus decoder input around the most important retrieved sentences.
How it connects
Connects FiD, MGFiD, passage reranking, context pruning, decoder fusion, and evidence grounding.

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Retrieval Augmented Generation, The Foundational Ideas. The section links point to the parts of the book it draws on.