# ROME: Memorization Insights from Text, Logits and Representation

arXiv: https://arxiv.org/abs/2403.00510 · HTML: https://mlsanity.com/paper/2403.00510/

## Key facts

- Cited by 8 papers in the map (data through 14 Oct 2026).
- Builds on 22 of its 28 references that are in the map.
- First posted on arXiv 1 Mar 2024.

## Builds on (most cited first, 22 of 22)

- [Language Models are Few-Shot Learners](https://mlsanity.com/paper/2005.14165/) (2020)
- [Llama 2: Open Foundation and Fine-Tuned Chat Models](https://mlsanity.com/paper/2307.09288/) (2023)
- [Evaluating Large Language Models Trained on Code](https://mlsanity.com/paper/2107.03374/) (2021)
- [Qwen Technical Report](https://mlsanity.com/paper/2309.16609/) (2023)
- [Extracting Training Data from Large Language Models](https://mlsanity.com/paper/2012.07805/) (2020)
- [Language Models as Knowledge Bases?](https://mlsanity.com/paper/1909.01066/) (2019)
- [When Not to Trust Language Models: Investigating Effectiveness of Parametric and Non-Parametric Memories](https://mlsanity.com/paper/2212.10511/) (2022)
- [The Secret Sharer: Evaluating and Testing Unintended Memorization in Neural Networks](https://mlsanity.com/paper/1802.08232/) (2018)
- [Deduplicating Training Data Makes Language Models Better](https://mlsanity.com/paper/2107.06499/) (2021)
- [What Neural Networks Memorize and Why: Discovering the Long Tail via Influence Estimation](https://mlsanity.com/paper/2008.03703/) (2020)
- [The Reversal Curse: LLMs trained on "A is B" fail to learn "B is A"](https://mlsanity.com/paper/2309.12288/) (2023)
- [Deduplicating Training Data Mitigates Privacy Risks in Language Models](https://mlsanity.com/paper/2202.06539/) (2022)
- [Chain-of-Thought Reasoning Without Prompting](https://mlsanity.com/paper/2402.10200/) (2024)
- [Counterfactual Memorization in Neural Language Models](https://mlsanity.com/paper/2112.12938/) (2021)
- [Scaling Laws and Interpretability of Learning from Repeated Data](https://mlsanity.com/paper/2205.10487/) (2022)
- [Multilingual LAMA: Investigating Knowledge in Multilingual Pretrained Language Models](https://mlsanity.com/paper/2102.00894/) (2021)
- [Understanding Transformer Memorization Recall Through Idioms](https://mlsanity.com/paper/2210.03588/) (2022)
- [On Memorization in Probabilistic Deep Generative Models](https://mlsanity.com/paper/2106.03216/) (2021)
- [Decoupling Knowledge from Memorization: Retrieval-augmented Prompt Learning](https://mlsanity.com/paper/2205.14704/) (2022)
- [How BPE Affects Memorization in Transformers](https://mlsanity.com/paper/2110.02782/) (2021)
- [An Empirical Study of Memorization in NLP](https://mlsanity.com/paper/2203.12171/) (2022)
- [Investigating Memorization of Conspiracy Theories in Text Generation](https://mlsanity.com/paper/2101.00379/) (2021)

## Built on by (most cited first, 8 of 8)

- [Beyond Math: Stories as a Testbed for Memorization-Constrained Reasoning in LLMs](https://mlsanity.com/paper/2412.14368/) (2024)
- [A Multi-Perspective Analysis of Memorization in Large Language Models](https://mlsanity.com/paper/2405.11577/) (2024)
- [Shared Path: Unraveling Memorization in Multilingual LLMs through Language Similarities](https://mlsanity.com/paper/2505.15722/) (2025)
- [Do LLMs Really Memorize Personally Identifiable Information? Revisiting PII Leakage with a Cue-Controlled Memorization Framework](https://mlsanity.com/paper/2601.03791/) (2026)
- [Characterizing Memorization in Diffusion Language Models: Generalized Extraction and Sampling Effects](https://mlsanity.com/paper/2603.02333/) (2026)
- [Memorization Inheritance in Sequence-Level Knowledge Distillation for Neural Machine Translation](https://mlsanity.com/paper/2502.01491/) (2025)
- [How Chain-of-Thought Works? Tracing Information Flow from Decoding, Projection, and Activation](https://mlsanity.com/paper/2507.20758/) (2025)
- [Causal Understanding by LLMs: The Role of Uncertainty](https://mlsanity.com/paper/2509.20088/) (2025)

## Abstract (arXiv)

Previous works have evaluated memorization by comparing model outputs with training corpora, examining how factors such as data duplication, model size, and prompt length influence memorization. However, analyzing these extensive training corpora is highly time-consuming. To address this challenge, this paper proposes an innovative approach named ROME that bypasses direct processing of the training data. Specifically, we select datasets categorized into three distinct types -- context-independent, conventional, and factual -- and redefine memorization as the ability to produce correct answers under these conditions. Our analysis then focuses on disparities between memorized and non-memorized samples by examining the logits and representations of generated texts. Experimental findings reveal that longer words are less likely to be memorized, higher confidence correlates with greater memorization, and representations of the same concepts are more similar across different contexts. Our code and data will be publicly available when the paper is accepted.