# Simpler becomes Harder: Do LLMs Exhibit a Coherent Behavior on Simplified Corpora?

arXiv: https://arxiv.org/abs/2404.06838 · HTML: https://mlsanity.com/paper/2404.06838/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 11 of its 29 references that are in the map.
- First posted on arXiv 10 Apr 2024.

## Builds on (most cited first, 11 of 11)

- [Adversarial Examples for Evaluating Reading Comprehension Systems](https://mlsanity.com/paper/1707.07328/) (2017)
- [Measuring and Improving Consistency in Pretrained Language Models](https://mlsanity.com/paper/2102.01017/) (2021)
- [Mind the Style of Text! Adversarial and Backdoor Attacks Based on Text Style Transfer](https://mlsanity.com/paper/2110.07139/) (2021)
- [Using Adversarial Attacks to Reveal the Statistical Bias in Machine Reading Comprehension Models](https://mlsanity.com/paper/2105.11136/) (2021)
- [Revisiting non-English Text Simplification: A Unified Multilingual Benchmark](https://mlsanity.com/paper/2305.15678/) (2023)
- [Searching for a Search Method: Benchmarking Search Algorithms for Generating NLP Adversarial Examples](https://mlsanity.com/paper/2009.06368/) (2020)
- [How do humans perceive adversarial text? A reality check on the validity and naturalness of word-based adversarial attacks](https://mlsanity.com/paper/2305.15587/) (2023)
- [Subjective Assessment of Text Complexity: A Dataset for German Language](https://mlsanity.com/paper/1904.07733/) (2019)
- [DEPLAIN: A German Parallel Corpus with Intralingual Translations into Plain Language for Sentence and Document Simplification](https://mlsanity.com/paper/2305.18939/) (2023)
- [How May I Help You? Using Neural Text Simplification to Improve Downstream NLP Tasks](https://mlsanity.com/paper/2109.04604/) (2021)
- [Evaluating Paraphrastic Robustness in Textual Entailment Models](https://mlsanity.com/paper/2306.16722/) (2023)

## Built on by (most cited first, 1 of 1)

- [Simplifications are Absolutists: How Simplified Language Reduces Word Sense Awareness in LLM-Generated Definitions](https://mlsanity.com/paper/2507.11981/) (2025)

## Abstract (arXiv)

Text simplification seeks to improve readability while retaining the original content and meaning. Our study investigates whether pre-trained classifiers also maintain such coherence by comparing their predictions on both original and simplified inputs. We conduct experiments using 11 pre-trained models, including BERT and OpenAI's GPT 3.5, across six datasets spanning three languages. Additionally, we conduct a detailed analysis of the correlation between prediction change rates and simplification types/strengths. Our findings reveal alarming inconsistencies across all languages and models. If not promptly addressed, simplified inputs can be easily exploited to craft zero-iteration model-agnostic adversarial attacks with success rates of up to 50%