LLM-Microscope: Uncovering the Hidden Role of Punctuation in Context Memory of Transformers (2025.findings-naacl)
Copied to clipboard
Anton Razzhigaev, Matvey Mikhalchuk, Temurbek Rahmatullaev, Elizaveta Goncharova, Polina Druzhinina, Ivan Oseledets, Andrey Kuznetsov
| Challenge: | Large Language Models (LLMs) encode and store contextual information, but internal mechanisms are opaque. |
| Approach: | They propose a toolkit that assesses token-level nonlinearity, evaluates contextual memory, visualizes intermediate layer contributions and measures intrinsic dimensionality of representations. |
| Outcome: | The proposed framework assesses token-level nonlinearity, evaluates contextual memory, visualizes intermediate layer contributions, and measures the intrinsic dimensionality of representations. |
Similar Papers
Punctuations and Predicates in Language Models (2026.findings-eacl)
Copied to clipboard
| Challenge: | Recent work has shown that LLMs perform tasks in ways that diverge significantly from human reasoning. |
| Approach: | They examine the computational importance of punctuation tokens in large language models . they use zeroing and layer-swapping techniques to examine their necessity and sufficiency . |
| Outcome: | The proposed model differs in GPT-2, DeepSeek, and Gemma in that punctuation is necessary and sufficient in multiple layers . the findings offer new insight into the internal mechanisms of punctuations in LLMs and have implications for interpretability and model analysis. |
Token Erasure as a Footprint of Implicit Vocabulary Items in LLMs (2024.emnlp-main)
Copied to clipboard
| Challenge: | Current language models process text as sequences of tokens that roughly correspond to words . individual tokens are often semantically unrelated to the meanings of the words/concepts they comprise . |
| Approach: | They propose a method to "read out" the implicit vocabulary of an autoregressive LLM by examining differences in token representations across layers. |
| Outcome: | The proposed method "reads out" the implicit vocabulary of an autoregressive LLM by examining differences in token representations across layers. |
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. |
Spelling-out is not Straightforward: LLMs’ Capability of Tokenization from Token to Characters (2025.findings-emnlp)
Copied to clipboard
| Challenge: | Large language models (LLMs) can spell out tokens character by character with high accuracy, yet struggle with more complex character-level tasks. |
| Approach: | They examine how large language models internally represent character-level information during the spelling-out process. |
| Outcome: | The embedding layer does not fully encode character-level information, especially beyond the first character. |
How Large Language Models Encode Context Knowledge? A Layer-Wise Probing Study (2024.lrec-main)
Copied to clipboard
| Challenge: | Existing studies have focused on enhancing the factualness of large language models using context knowledge. |
| Approach: | They propose to use ChatGPT to construct probing datasets that provide diverse and coherent evidence corresponding to various facts. |
| Outcome: | The proposed model can encode knowledge across different layers, and it is compared with existing models. |
Frozen LLMs are Native Decoders for High-Norm Semantic Vectors (2026.acl-long)
Copied to clipboard
| Challenge: | Existing compression methods selectively prune tokens based on information-theoretic metrics, resulting in interpretability but risking the loss of fine-grained information. |
| Approach: | They propose a landmark-based compression framework for long contexts that captures global dependencies over landmark tokens. |
| Outcome: | The proposed framework outperforms soft compression baselines on four QA benchmarks. |
Insights into LLM Long-Context Failures: When Transformers Know but Don’t Tell (2024.findings-emnlp)
Copied to clipboard
| Challenge: | Large Language Models (LLMs) exhibit positional bias, struggling to utilize information from the middle or end of long contexts. |
| Approach: | They propose to examine LLMs' long-context generalizations by probing their hidden representations. |
| Outcome: | The proposed models excel at processing extended contexts while preserving their positional bias. |
Token-wise Decomposition of Autoregressive Language Model Hidden States for Analyzing Model Predictions (2023.acl-long)
Copied to clipboard
| Challenge: | Recent work on why Transformer-based large language models make predictions has made their behavior opaque due to the complexity of the computations performed within each layer. |
| Approach: | They propose a linear decomposition of final hidden states from autoregressive language models based on each initial input token, which is exact for virtually all contemporary Transformer architectures. |
| Outcome: | The proposed method analyzes the influence of input tokens on model probabilities over a sequence of upcoming words with only one forward pass from the model. |
Exploring Context Strategies in LLMs for Discourse-Aware Machine Translation (2025.findings-emnlp)
Copied to clipboard
| Challenge: | Large language models excel at machine translation, but the impact of how LLMs utilize different forms of contextual information on discourse-level phenomena remains underexplored. |
| Approach: | They examine how different forms of context influence standard MT metrics and specific discourse phenomena such as formality, pronoun selection, and lexical cohesion. |
| Outcome: | Evaluating multiple LLMs across multiple domains and language pairs, the findings consistently show that context boosts translation and discourse-specific performance. |
Can Large Language Models Understand Context? (2024.findings-eacl)
Copied to clipboard
Yilun Zhu, Joel Moniz, Shruti Bhargava, Jiarui Lu, Dhivya Piraviperumal, Site Li, Yuan Zhang, Hong Yu, Bo-Hsiang Tseng
| Challenge: | Existing evaluation methodologies for Large Language Models (LLMs) have been inadequate to evaluate their ability to understand contextual features. |
| Approach: | They propose a benchmark to assess large language models' ability to understand context by adapting existing datasets to suit their evaluation. |
| Outcome: | The proposed model performs better under the in-context learning pretraining scenario than state-of-the-art models. |