Impact of Task Phrasing on Presumptions in Large Language Models

Kenneth J. K. Ong · 2026

Read Impact of Task Phrasing on Presumptions in Large Language Models by Kenneth J. K. Ong (2026) 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-2605-00436.txt.

Description

Kenneth J. K. Ong studies how the phrasing of a task can plant presumptions in a large language model, quietly bending its decisions. In unpredictable real-world use, such presumptions are a safety and reliability concern: a model locked into an assumed framing adapts badly when the real task deviates. Ong tests the effect with a classic game as the case study.

In an iterated prisoner's dilemma, the models proved susceptible to presumptions in their choices, even when they reasoned step by step before deciding. When the task phrasing was kept neutral, the same models reasoned logically with little sign of the bias. The contrast is the paper's lesson. Careful, neutral task phrasing is not cosmetic; it measurably reduces the risk that a model will answer the task it expected instead of the one it was given.

Catalog record

CREATORKenneth J. K. Ong
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE19 KB
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2605.00436

Browse the full library · About the Tower of Babel