# Retriever-Free Retrieval: Navigating Documents with Decision Models

Source: [arXiv](https://arxiv.org/abs/2610.35868v1)  
Feed7 permalink: https://feed7.dev/p/2610-35868v1-1uzxrmk  
Published: 2026-10-26T17:59:44.000Z  
Trust: Needs Review (needs_review)

## Why Included

DecisionRAG uses an LLM to navigate document branches without a first-stage retriever. Its reported gains suggest retrieval quality can improve by changing how candidates are searched, not only how they are ranked.

## Source Summary

DecisionRAG uses a language model to choose which branches of a corpus to explore, then narrows from broad topics to passages. It generates no tokens during these decisions and requires no task-specific training or first-stage retriever. Across four benchmarks, it ran in **0.7–1.8 seconds** per query, including on **2.68 million** Wikipedia paragraphs.

## Practical Implication

If your agent depends on retrieval, this points to a possible alternative to lexical or embedding search followed by reranking: use a decision model to guide corpus navigation. The reported comparison beat dense retrieval with an 8B reranker by **13.8 R@1 points** on MultiDoc2Dial and **17.4 Hit@1 points** on Natural Questions.

## Agent-Ready Context

DecisionRAG uses a language model to choose which branches of a corpus to explore, then narrows from broad topics to passages. It generates no tokens during these decisions and requires no task-specific training or first-stage retriever. Across four benchmarks, it ran in **0.7–1.8 seconds** per query, including on **2.68 million** Wikipedia paragraphs.

If your agent depends on retrieval, this points to a possible alternative to lexical or embedding search followed by reranking: use a decision model to guide corpus navigation. The reported comparison beat dense retrieval with an 8B reranker by **13.8 R@1 points** on MultiDoc2Dial and **17.4 Hit@1 points** on Natural Questions.

These are paper results, and the material does not describe a production implementation or its deployment costs. It remains open how well this approach transfers to other corpora and workloads.

## Connected Context

Feed7 judgment across 928 accumulated Signals:

DecisionRAG presents corpus navigation as an alternative retrieval route: a model selects branches and passages without a first-stage retriever. Its reported benchmark results make it a candidate to compare with existing pipelines, while the supplied work on chunking, memory, and persistent knowledge shows that retrieval quality and cost remain workload- and architecture-dependent.

- [Stop Chunking Like It's 2022 — Yuval Belfer, AI21 Labs](https://feed7.dev/p/stop-chunking-like-it-s-2022-yuval-belfer-ai21-labs-1bsztja) — Both challenge a fixed retrieval pipeline: DecisionRAG replaces the first-stage retriever with guided navigation, while multiscale indexing changes chunk granularity and fusion.
- [Graphify-Labs/graphify](https://feed7.dev/p/graphify-1e0bs1f) — Graphify exposes repository relationships as a queryable graph; DecisionRAG’s branch navigation offers a different way to narrow a corpus before selecting relevant evidence.
- [nashsu/llm_wiki](https://feed7.dev/p/llm-wiki-1dqmknq) — LLM Wiki compiles sources into persistent, traceable pages, whereas DecisionRAG navigates the corpus at query time; the approaches differ in when and how retrieved context is organized.
- [Structured Memory for Edge Language Models: Persistent Context and Corpus Retrieval via O(1) SSM State Injection](https://feed7.dev/p/2608-02560v1-1u8ls14) — PRECOG stores reusable corpus states for compatible SSMs to reduce repeated ingestion, while DecisionRAG chooses passages through query-time navigation; their gains depend on different model and deployment constraints.

## Context Map

- Layer: context
- Domains: research, data
- Topics: retrieval, context-engineering

## Uncertainty

- These are paper results, and the material does not describe a production implementation or its deployment costs. It remains open how well this approach transfers to other corpora and workloads.

## Agent Instruction

Use this item as source-backed context. Do not invent claims beyond the linked source. If this item conflicts with another source, call out the conflict.