# Towards Human Understanding of Paraphrase Types in Large Language Models

arXiv: https://arxiv.org/abs/2407.02302 · HTML: https://mlsanity.com/paper/2407.02302/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Placed in Human values and AI feedback alignment, one of 10 directions in LLM alignment / RLHF / preference optimisation.
- Builds on 11 of its 34 references that are in the map.
- First posted on arXiv 2 Jul 2024.

## Placed in

- LLM alignment / RLHF / preference optimisation / [Human values and AI feedback alignment](https://mlsanity.com/alignment/human-values-and-ai-feedback-alignment/)

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

- [Training language models to follow instructions with human feedback](https://mlsanity.com/paper/2203.02155/) (2022)
- [Llama 2: Open Foundation and Fine-Tuned Chat Models](https://mlsanity.com/paper/2307.09288/) (2023)
- [Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedback](https://mlsanity.com/paper/2204.05862/) (2022)
- [Large Language Models Are Human-Level Prompt Engineers](https://mlsanity.com/paper/2211.01910/) (2022)
- [Open Problems and Fundamental Limitations of Reinforcement Learning from Human Feedback](https://mlsanity.com/paper/2307.15217/) (2023)
- [Is ChatGPT better than Human Annotators? Potential and Limitations of ChatGPT in Explaining Implicit Hate Speech](https://mlsanity.com/paper/2302.07736/) (2023)
- [FLEX: Unifying Evaluation for Few-Shot NLP](https://mlsanity.com/paper/2107.07170/) (2021)
- [Generating Syntactically Controlled Paraphrases without Using Annotated Parallel Pairs](https://mlsanity.com/paper/2101.10579/) (2021)
- [How Large Language Models are Transforming Machine-Paraphrased Plagiarism](https://mlsanity.com/paper/2210.03568/) (2022)
- [Paraphrase Types Elicit Prompt Engineering Capabilities](https://mlsanity.com/paper/2406.19898/) (2024)
- [Paraphrase Identification with Deep Learning: A Review of Datasets and Methods](https://mlsanity.com/paper/2212.06933/) (2022)

## Built on by (most cited first, 2 of 2)

- [Say It Another Way: Auditing LLMs with a User-Grounded Automated Paraphrasing Framework](https://mlsanity.com/paper/2505.03563/) (2025)
- [Same Question, Different Answers: Evaluating LLM Reliability Beyond Accuracy](https://mlsanity.com/paper/2607.22554/) (2026)

## Abstract (arXiv)

Paraphrases represent a human's intuitive ability to understand expressions presented in various different ways. Current paraphrase evaluations of language models primarily use binary approaches, offering limited interpretability of specific text changes. Atomic paraphrase types (APT) decompose paraphrases into different linguistic changes and offer a granular view of the flexibility in linguistic expression (e.g., a shift in syntax or vocabulary used). In this study, we assess the human preferences towards ChatGPT in generating English paraphrases with ten APTs and five prompting techniques. We introduce APTY (Atomic Paraphrase TYpes), a dataset of 800 sentence-level and word-level annotations by 15 annotators. The dataset also provides a human preference ranking of paraphrases with different types that can be used to fine-tune models with RLHF and DPO methods. Our results reveal that ChatGPT and a DPO-trained LLama 7B model can generate simple APTs, such as additions and deletions, but struggle with complex structures (e.g., subordination changes). This study contributes to understanding which aspects of paraphrasing language models have already succeeded at understanding and what remains elusive. In addition, we show how our curated datasets can be used to develop language models with specific linguistic capabilities.