Hallucination Behavior in Multimodal LLMs Across Agricultural Image Interpretation and Generation Tasks

Partho Ghose, Al Bashir, Prem Raj, Azlan Zahid · 2026

Read Hallucination Behavior in Multimodal LLMs Across Agricultural Image Interpretation and Generation Tasks by Partho Ghose, Al Bashir, Prem Raj, Azlan Zahid (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-2605-27595.txt.

Description

Partho Ghose, Al Bashir, Prem Raj, and Azlan Zahid study hallucinations in multimodal models for agricultural imaging. In image-to-text tasks, zero-shot accuracy runs only 63 to 75 percent, while few-shot prompting lifts it to about 87 percent, though false detections and missed infections persist. In text-to-image generation, relaxed prompts yield up to 91 percent biologically inconsistent scenes, exposing deep reliability gaps.

Catalog record

CREATORPartho Ghose, Al Bashir, Prem Raj, Azlan Zahid
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE35562
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2605.27595

Browse the full library · About the Tower of Babel