Retrieval as Generation: A Unified Framework with Self-Triggered Information Planning

TL;DR

This paper introduces GRIP, a unified retrieval-as-generation framework that integrates retrieval control into token-level decoding, enabling dynamic, end-to-end reasoning and evidence gathering within a single model.

Contribution

It proposes a novel approach where retrieval decisions are embedded within generation, allowing for flexible, multi-step inference without external retrieval controllers.

Findings

GRIP outperforms strong RAG baselines on five QA benchmarks.

It is competitive with GPT-4o while using fewer parameters.

The framework enables dynamic, on-the-fly evidence integration.

Abstract

We revisit retrieval-augmented generation (RAG) by embedding retrieval control directly into generation. Instead of treating retrieval as an external intervention, we express retrieval decisions within token-level decoding, enabling end-to-end coordination without additional controllers or classifiers. Under the paradigm of Retrieval as Generation, we propose \textbf{GRIP} (\textbf{G}eneration-guided \textbf{R}etrieval with \textbf{I}nformation \textbf{P}lanning), a unified framework in which the model regulates retrieval behavior through control-token emission. Central to GRIP is \textit{Self-Triggered Information Planning}, which allows the model to decide when to retrieve, how to reformulate queries, and when to terminate, all within a single autoregressive trajectory. This design tightly couples retrieval and reasoning and supports dynamic multi-step inference with on-the-fly evidence integration. To supervise these behaviors, we construct a structured training set covering answerable, partially answerable, and multi-hop queries, each aligned with specific token patterns. Experiments on five QA benchmarks show that GRIP surpasses strong RAG baselines and is competitive with GPT-4o while using substantially fewer parameters.