SUGAR: Leveraging Contextual Confidence for Smarter Retrieval

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Zubkova, Hanna, Park, Ji-Hoon, Lee, Seong-Whan
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912181851783168
author Zubkova, Hanna
Park, Ji-Hoon
Lee, Seong-Whan
author_facet Zubkova, Hanna
Park, Ji-Hoon
Lee, Seong-Whan
contents Bearing in mind the limited parametric knowledge of Large Language Models (LLMs), retrieval-augmented generation (RAG) which supplies them with the relevant external knowledge has served as an approach to mitigate the issue of hallucinations to a certain extent. However, uniformly retrieving supporting context makes response generation source-inefficient, as triggering the retriever is not always necessary, or even inaccurate, when a model gets distracted by noisy retrieved content and produces an unhelpful answer. Motivated by these issues, we introduce Semantic Uncertainty Guided Adaptive Retrieval (SUGAR), where we leverage context-based entropy to actively decide whether to retrieve and to further determine between single-step and multi-step retrieval. Our empirical results show that selective retrieval guided by semantic uncertainty estimation improves the performance across diverse question answering tasks, as well as achieves a more efficient inference.
format Preprint
id arxiv_https___arxiv_org_abs_2501_04899
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle SUGAR: Leveraging Contextual Confidence for Smarter Retrieval
Zubkova, Hanna
Park, Ji-Hoon
Lee, Seong-Whan
Computation and Language
Artificial Intelligence
Bearing in mind the limited parametric knowledge of Large Language Models (LLMs), retrieval-augmented generation (RAG) which supplies them with the relevant external knowledge has served as an approach to mitigate the issue of hallucinations to a certain extent. However, uniformly retrieving supporting context makes response generation source-inefficient, as triggering the retriever is not always necessary, or even inaccurate, when a model gets distracted by noisy retrieved content and produces an unhelpful answer. Motivated by these issues, we introduce Semantic Uncertainty Guided Adaptive Retrieval (SUGAR), where we leverage context-based entropy to actively decide whether to retrieve and to further determine between single-step and multi-step retrieval. Our empirical results show that selective retrieval guided by semantic uncertainty estimation improves the performance across diverse question answering tasks, as well as achieves a more efficient inference.
title SUGAR: Leveraging Contextual Confidence for Smarter Retrieval
topic Computation and Language
Artificial Intelligence
url https://arxiv.org/abs/2501.04899