| Challenge: | Existing methods for analyzing memorization use definitions that are based on model performance, which changes between models and often also between training runs. |
| Approach: | They propose idioms as inputs that typically trigger memory recall and propose a set of English idiomas to test their methodological framework for probing and characterizing recall of memorized sequences in transformer LMs. |
| Outcome: | The proposed framework compares model behavior on memorized vs. non-memorized inputs across different model sizes and architectures. |
Similar Papers
Memorization or Reasoning? Exploring the Idiom Understanding of LLMs (2025.emnlp-main)
Copied to clipboard
| Challenge: | idioms have long posed a challenge due to their unique linguistic properties, which set them apart from other common expressions. |
| Approach: | They propose to use a large-scale dataset of idioms in six languages to evaluate LLMs' idiomatic processing ability. |
| Outcome: | The proposed model integrates contextual cues and reasoning to improve idiom understanding in LLMs, suggesting that their performance is influenced by memorization and reasoning. |
A Multi-Perspective Analysis of Memorization in Large Language Models (2024.emnlp-main)
Copied to clipboard
| Challenge: | Large Language Models (LLMs) can generate the same sequences contained in the pre-train corpus, known as memorization. |
| Approach: | They analyze the relationship between memorization and outputs from Large Language Models (LLMs) they show a sudden drop and increase in the frequency of input tokens when generating memorized/unmemorized sequences . |
| Outcome: | The proposed model can generate the same sequences contained in the pre-train corpus, and it can predict unmemorized tokens. |
Unlocking Memorization in Large Language Models with Dynamic Soft Prompting (2024.emnlp-main)
Copied to clipboard
Zhepeng Wang, Runxue Bao, Yawen Wu, Jackson Taylor, Cao Xiao, Feng Zheng, Weiwen Jiang, Shangqian Gao, Yanfu Zhang
| Challenge: | Pretrained large language models excel in a variety of natural language processing tasks . however, they pose significant security risks due to their tendency to memorize training data . |
| Approach: | They propose a method to estimate LLM memorization using dynamic, prefix-dependent soft prompts. |
| Outcome: | The proposed method can achieve maximum relative improvement of 135.3% and 39.8% over baseline compared to state-of-the-art methods. |
Position Paper: MeMo: Towards Language Models with Associative Memory Mechanisms (2025.findings-acl)
Copied to clipboard
Fabio Massimo Zanzotto, Elena Sofia Ruzzetti, Giancarlo A. Xompero, Leonardo Ranaldi, Davide Venditti, Federico Ranaldi, Cristina Giannone, Andrea Favalli, Raniero Romagnoli
| Challenge: | Memorization is a fundamental ability of Transformer-based Large Language Models, achieved through learning. |
| Approach: | They propose an architecture that explicitly memorizes sequences of tokens in layered associative memories. |
| Outcome: | The proposed architecture shows that memorization is a fundamental ability of large language models, achieved through learning. |
What Context Features Can Transformer Language Models Use? (2021.acl-long)
Copied to clipboard
| Challenge: | Recent studies show that transformer-based language models benefit from conditioning on contexts of hundreds to thousands of previous tokens. |
| Approach: | They propose to use lexical and structural information to ablate usable information in transformer language models. |
| Outcome: | The proposed model improves when conditioning on contexts of thousands of previous tokens. |
VISIT: Visualizing and Interpreting the Semantic Information Flow of Transformers (2023.findings-emnlp)
Copied to clipboard
| Challenge: | Recent work in interpretability suggests we can project weights and hidden states of transformer-based language models (LMs) to their vocabulary space, a transformation that makes them more human interpretable. |
| Approach: | They propose a tool to visualize a forward pass of Generative Pre-trained Transformers as an interactive flow graph with nodes representing neurons or hidden states and edges representing interactions between them. |
| Outcome: | The proposed visualization simplifies huge amounts of data into easy-to-read graphs that can reflect the models’ internal processing, uncovering the contribution of each component to the models' final prediction. |
Transformers Can Represent n-gram Language Models (2024.naacl-long)
Copied to clipboard
| Challenge: | Existing work studies the transformer architecture in terms of language acceptance . n-gram LMs can represent any ngram lm, giving a concrete lower bound on their probabilistic representational capacity. |
| Approach: | They propose to use transformer LMs to represent n-gram LM, a class of language models . they show transformer lms can exactly represent any ngram lmm, giving them a lower bound . |
| Outcome: | The proposed model can represent any n-gram LM, giving a lower bound on their probabilistic representational capacity. |
Can Transformers Learn n-gram Language Models? (2024.emnlp-main)
Copied to clipboard
| Challenge: | Existing work has tested transformers' ability to represent formal languages, but language models are not classifiers of strings but rather distributions over them. |
| Approach: | They relate transformers' ability to learn random n-gram language models to ngram language model (LM) they find add- smoothing outperforms transformers on the former, while transformers perform better on the latter . |
| Outcome: | The proposed models outperform classical methods designed to learn n-gram LMs, while transformers perform better on the latter. |
Dissecting Recall of Factual Associations in Auto-Regressive Language Models (2023.emnlp-main)
Copied to clipboard
| Challenge: | Existing studies have focused on identifying where factual knowledge is encoded in the network, but little is known about how it is extracted from the model parameters during inference. |
| Approach: | They examine how factual associations are stored and retrieved internally in LMs . they use attention edges to identify critical points where information propagates to the prediction . |
| Outcome: | The proposed model aggregates information about subject and relation to predict the correct attribute . the model “queries” the enriched subject to extract the attribute based on the proposed model . |
Do transformer models do phonology like a linguist? (2023.findings-acl)
Copied to clipboard
| Challenge: | Neural sequence-to-sequence models have been very successful at tasks in phonology and morphology that require a capacity for intricate linguistic generalisations. |
| Approach: | They propose to use a transformer model to model phonological generalisations on 29 different processes . they also use phonology as a primer and a rule with a withheld case to infer behaviour . |
| Outcome: | The proposed model can model all 29 phonological phenomena on a constructed language, regardless of perceived process difficulty. |