Understanding Subword Compositionality of Large Language Models (2025.emnlp-main)
Copied to clipboard
| Challenge: | Large language models (LLMs) take sequences of subwords as input, requiring them to compose subword representations into meaningful word-level representations. |
| Approach: | They propose to probe how large language models compose subword information . they find structural similarity, semantic decomposability, and form retention are key aspects . |
| Outcome: | The proposed models can be classified into three distinct groups, the authors show . they show that they can achieve great performance when probing layer by layer their sensitivity to semantic decompositionality . |
Similar Papers
A Systematic Study of Leveraging Subword Information for Learning Word Representations (N19-1)
Copied to clipboard
| Challenge: | Existing word representation models for morphologically rich languages use subword-level information, but their systematic comparative analysis across typologically diverse languages and tasks is still missing. |
| Approach: | They propose a framework for learning subword-informed word representations that allows for easy experimentation with different segmentation and composition components. |
| Outcome: | The proposed framework allows for easy experimentation with different segmentation and composition components, as well as advanced techniques based on position embeddings and self-attention. |
Are representations built from the ground up? An empirical examination of local composition in language models (2022.emnlp-main)
Copied to clipboard
| Challenge: | Compositionality is a hallmark of human language, but many phrases are non-compositional . a study by a team of researchers shows that LMs may not be able to distinguish between compositional and non-composable phrases. |
| Approach: | They propose to predict LM-internal representations of longer phrases given their constituents . they find that the representation of a parent phrase can be predicted with some accuracy . |
| Outcome: | The proposed model can predict a parent phrase with some accuracy given its children's transformations, but this is not the case. |
Is a cute puyfred cute? Context-dependent form-meaning systematicity in LLMs (2025.findings-acl)
Copied to clipboard
| Challenge: | valence is encoded in meaningful ways in large language models and in some LLMs, pseudowords affect the representation of whole sentences similarly to words. |
| Approach: | They investigate how LLMs represent valence, a key semantic attribute, and how they deal with contextualisation of pseudowords in sentences. |
| Outcome: | The results show that the models represent valence, a key semantic attribute, in sentences and in context, and that they handle the contextualisation of pseudowords differently. |
How Do Large Language Models Capture the Ever-changing World Knowledge? A Review of Recent Advances (2023.emnlp-main)
Copied to clipboard
| Challenge: | Large language models (LLMs) are impressive in solving tasks, but they can quickly be outdated after deployment. |
| Approach: | They provide a review of recent advances in aligning deployed large language models with the ever-changing world knowledge. |
| Outcome: | The proposed models can be used to perform various tasks directly through in-context learning or for further fine-tuning for domain-specific uses. |
Unlike “Likely”, “Unlike” is Unlikely: BPE-based Segmentation hurts Morphological Derivations in LLMs (2025.coling-main)
Copied to clipboard
| Challenge: | Large Language Models (LLMs) use subword vocabularies to process and generate text. |
| Approach: | They find that Large Language Models (LLMs) perform poorly at handling some types of affixations because subwords are marked as initial- or intra-word . |
| Outcome: | The largest models trained on enough data can mitigate this tendency because initial- and intra-word embeddings are aligned; in-context learning also helps when all examples are selected in a consistent way; but only morphological segmentation can achieve a near-perfect accuracy. |
Where Do LLMs Compose Meaning? A Layerwise Analysis of Compositional Robustness (2026.eacl-long)
Copied to clipboard
| Challenge: | Existing approaches to linguistic compositional processing are limited by hallucinations, limited interpretability and a lack of semantic transparency. |
| Approach: | They propose a method that pooles token representations into linguistic constituents at various layers to improve model reliability and interpretability. |
| Outcome: | Experiments on inverse definition modelling, hypernym and synonym prediction show that semantic composition is not localised to specific layers but distributed across network depth. |
What Does BERT Learn about the Structure of Language? (P19-1)
Copied to clipboard
| Challenge: | BERT is a language representation model that has performed well in diverse language understanding benchmarks. |
| Approach: | They perform experiments to unpack the elements of English language structure learned by BERT. |
| Outcome: | The proposed model outperforms state-of-the-art models in the GLUE benchmark by a significant margin. |
Large Vocabulary Size Improves Large Language Models (2025.findings-acl)
Copied to clipboard
| Challenge: | Existing studies have investigated the properties of internal layers in large language models, but no studies have defined the vocabulary size. |
| Approach: | They propose a method to use a new vocabulary instead of the pre-defined one in a continual training scenario. |
| Outcome: | The proposed method outperforms the model with the pre-defined vocabulary in a continual training scenario. |
Subword Segmentation in LLMs: Looking at Inflection and Consistency (2024.emnlp-main)
Copied to clipboard
| Challenge: | Subword segmentation is not linguistically guided and is not currently well understood in LLMs. |
| Approach: | They group words according to their segmentation properties and compare how well a model can solve a linguistic task for these groups using two criteria: adherence to morpheme boundaries and segmentation consistency of inflected forms of a lemma. |
| Outcome: | The results show that the criterion of segmentation consistency can predict the model’s ability to recognize and generate the lemma from an inflected form, providing evidence that subword segmentation is relevant. |
Representational Isomorphism and Alignment of Multilingual Large Language Models (2024.findings-emnlp)
Copied to clipboard
| Challenge: | Existing isomorphism of sentence representations can facilitate representational alignments in zero-shot and few-shot settings. |
| Approach: | They propose to apply a contrastive objective to LLMs with a small number of translation pairs to improve models' performance on Semantic Textual Similarity tasks. |
| Outcome: | The proposed representation-level approach significantly improves on Semantic Textual Similarity (STS) tasks across languages even without a monolingual objective. |