Prompting Large Language Models with Partial Knowledge for Answering Questions with Unseen Entities DOI: XXXXXXX.XXXXXXX ISBN: 978-1-4503-XXXX-X/2018/06 CCS: Information systems Question answering CCS: Computing methodologies Discourse, dialogue and pragmatics

Zhichao Yan, Jiapu Wang, Jiaoyan Chen, Yanyan Wang, Hongye Tan, Jiye Liang, Xiaoli Li, Ru Li, Jeff Z.Pan · 2025

Read Prompting Large Language Models with Partial Knowledge for Answering Questions with Unseen Entities DOI: XXXXXXX.XXXXXXX ISBN: 978-1-4503-XXXX-X/2018/06 CCS: Information systems Question answering CCS: Computing methodologies Discourse, dialogue and pragmatics by Zhichao Yan, Jiapu Wang, Jiaoyan Chen, Yanyan Wang, Hongye Tan, Jiye Liang, Xiaoli Li, Ru Li, Jeff Z.Pan (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2508-01290.txt.

Description

Yan and co-authors show in 2025 that partially relevant knowledge can awaken answers inside large language models for questions about unseen entities. Retrieval systems add outside knowledge, which splits into direct evidence, clues, and weak context. Weak context with partial relevance is usually treated as noise.

Their view flips that assumption, since models may already hold the missing link. Gold reasoning paths were trimmed to build partial knowledge, and theory plus tests on two graph question datasets back the effect. A new task mimics failed entity linking in incomplete graphs. Their approach beat embedding similarity methods in practical trials, which reframes partial retrieval as an asset.

Catalog record

CREATORZhichao Yan, Jiapu Wang, Jiaoyan Chen, Yanyan Wang, Hongye Tan, Jiye Liang, Xiaoli Li, Ru Li, Jeff Z.Pan
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE72 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2508-01290

Browse the full library · About the Tower of Babel