Saved in:
Bibliographic Details
Main Authors: Vasileva, Lisa, Sim, Karin
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2604.15165
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913038411497472
author Vasileva, Lisa
Sim, Karin
author_facet Vasileva, Lisa
Sim, Karin
contents LLMs are proving to be adept at machine translation although due to their generative nature they may at times overgenerate in various ways. These overgenerations are different from the neurobabble seen in NMT and range from LLM self-explanations, to risky confabulations, to appropriate explanations, where the LLM is able to act as a human translator would, enabling greater comprehension for the target audience. Detecting and determining the exact nature of the overgenerations is a challenging task. We detail different strategies we have explored for our work in a commercial setting, and present our results.
format Preprint
id arxiv_https___arxiv_org_abs_2604_15165
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Fabricator or dynamic translator?
Vasileva, Lisa
Sim, Karin
Computation and Language
LLMs are proving to be adept at machine translation although due to their generative nature they may at times overgenerate in various ways. These overgenerations are different from the neurobabble seen in NMT and range from LLM self-explanations, to risky confabulations, to appropriate explanations, where the LLM is able to act as a human translator would, enabling greater comprehension for the target audience. Detecting and determining the exact nature of the overgenerations is a challenging task. We detail different strategies we have explored for our work in a commercial setting, and present our results.
title Fabricator or dynamic translator?
topic Computation and Language
url https://arxiv.org/abs/2604.15165