Saved in:
Bibliographic Details
Main Authors: Huang, Tianyi, Yang, Caden, Yin, Emily, Wang, Eric, Zhang, Michael
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2603.20673
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914439065763840
author Huang, Tianyi
Yang, Caden
Yin, Emily
Wang, Eric
Zhang, Michael
author_facet Huang, Tianyi
Yang, Caden
Yin, Emily
Wang, Eric
Zhang, Michael
contents Retrieval-augmented language models can retrieve relevant evidence yet still commit to answers before explicitly checking whether the retrieved context supports the conclusion. We present PAVE (Premise-Grounded Answer Validation and Editing), an inference-time validation layer for evidence-grounded question answering. PAVE decomposes retrieved context into question-conditioned atomic facts, drafts an answer, scores how well that draft is supported by the extracted premises, and revises low-support outputs before finalization. The resulting trace makes answer commitment auditable at the level of explicit premises, support scores, and revision decisions. In controlled ablations with a fixed retriever and backbone, PAVE outperforms simpler post-retrieval baselines in two evidence-grounded QA settings, with the largest gain reaching 32.7 accuracy points on a span-grounded benchmark. We view these findings as proof-of-concept evidence that explicit premise extraction plus support-gated revision can strengthen evidence-grounded consistency in retrieval-augmented LLM systems.
format Preprint
id arxiv_https___arxiv_org_abs_2603_20673
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle PAVE: Premise-Aware Validation and Editing for Retrieval-Augmented LLMs
Huang, Tianyi
Yang, Caden
Yin, Emily
Wang, Eric
Zhang, Michael
Computation and Language
Artificial Intelligence
Retrieval-augmented language models can retrieve relevant evidence yet still commit to answers before explicitly checking whether the retrieved context supports the conclusion. We present PAVE (Premise-Grounded Answer Validation and Editing), an inference-time validation layer for evidence-grounded question answering. PAVE decomposes retrieved context into question-conditioned atomic facts, drafts an answer, scores how well that draft is supported by the extracted premises, and revises low-support outputs before finalization. The resulting trace makes answer commitment auditable at the level of explicit premises, support scores, and revision decisions. In controlled ablations with a fixed retriever and backbone, PAVE outperforms simpler post-retrieval baselines in two evidence-grounded QA settings, with the largest gain reaching 32.7 accuracy points on a span-grounded benchmark. We view these findings as proof-of-concept evidence that explicit premise extraction plus support-gated revision can strengthen evidence-grounded consistency in retrieval-augmented LLM systems.
title PAVE: Premise-Aware Validation and Editing for Retrieval-Augmented LLMs
topic Computation and Language
Artificial Intelligence
url https://arxiv.org/abs/2603.20673