# Discovering Semantic Subdimensions through Disentangled Conceptual Representations

arXiv: https://arxiv.org/abs/2508.21436 · HTML: https://mlsanity.com/paper/2508.21436/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 19 of its 82 references that are in the map.
- First posted on arXiv 29 Aug 2025.

## Builds on (most cited first, 19 of 19)

- [Efficient Estimation of Word Representations in Vector Space](https://mlsanity.com/paper/1301.3781/) (2013)
- [Adversarial Example Generation with Syntactically Controlled Paraphrase Networks](https://mlsanity.com/paper/1804.06059/) (2018)
- [Revisiting Pre-Trained Models for Chinese Natural Language Processing](https://mlsanity.com/paper/2004.13922/) (2020)
- [Variational Dropout Sparsifies Deep Neural Networks](https://mlsanity.com/paper/1701.05369/) (2017)
- [Towards Scientific Intelligence: A Survey of LLM-based Scientific Agents](https://mlsanity.com/paper/2503.24047/) (2025)
- [A Multi-Task Approach for Disentangling Syntax and Semantics in Sentence Representations](https://mlsanity.com/paper/1904.01173/) (2019)
- [Contrast, Attend and Diffuse to Decode High-Resolution Images from Brain Activities](https://mlsanity.com/paper/2305.17214/) (2023)
- [Contributions of Shape, Texture, and Color in Visual Recognition](https://mlsanity.com/paper/2207.09510/) (2022)
- [ChineseWebText: Large-scale High-quality Chinese Web Text Extracted with Effective Evaluation Model](https://mlsanity.com/paper/2311.01149/) (2023)
- [CROP: Contextual Region-Oriented Visual Token Pruning](https://mlsanity.com/paper/2505.21233/) (2025)
- [Learning Disentangled Representations for Natural Language Definitions](https://mlsanity.com/paper/2210.02898/) (2022)
- [Teaching Vision-Language Models to Ask: Resolving Ambiguity in Visual Questions](https://mlsanity.com/paper/2507.13773/) (2025)
- [MulCogBench: A Multi-modal Cognitive Benchmark Dataset for Evaluating Chinese and English Computational Language Models](https://mlsanity.com/paper/2403.01116/) (2024)
- [How Syntax Specialization Emerges in Language Models](https://mlsanity.com/paper/2505.19548/) (2025)
- [StyleFlow: Disentangle Latent Representations via Normalizing Flow for Unsupervised Text Style Transfer](https://mlsanity.com/paper/2212.09670/) (2022)
- [Computational Models to Study Language Processing in the Human Brain: A Survey](https://mlsanity.com/paper/2403.13368/) (2024)
- [Navigating Brain Language Representations: A Comparative Analysis of Neural Language Models and Psychologically Plausible Models](https://mlsanity.com/paper/2404.19364/) (2024)
- [Disentangling Linguistic Features with Dimension-Wise Analysis of Vector Embeddings](https://mlsanity.com/paper/2504.14766/) (2025)
- [Decoding the Multimodal Mind: Generalizable Brain-to-Text Translation via Multimodal Alignment and Adaptive Routing](https://mlsanity.com/paper/2505.10356/) (2025)

## Built on by (most cited first, 1 of 1)

- [Look Again, Think Slowly: Enhancing Visual Reflection in Vision-Language Models](https://mlsanity.com/paper/2509.12132/) (2025)

## Abstract (arXiv)

Understanding the core dimensions of conceptual semantics is fundamental to uncovering how meaning is organized in language and the brain. Existing approaches often rely on predefined semantic dimensions that offer only broad representations, overlooking finer conceptual distinctions. This paper proposes a novel framework to investigate the subdimensions underlying coarse-grained semantic dimensions. Specifically, we introduce a Disentangled Continuous Semantic Representation Model (DCSRM) that decomposes word embeddings from large language models into multiple sub-embeddings, each encoding specific semantic information. Using these sub-embeddings, we identify a set of interpretable semantic subdimensions. To assess their neural plausibility, we apply voxel-wise encoding models to map these subdimensions to brain activation. Our work offers more fine-grained interpretable semantic subdimensions of conceptual meaning. Further analyses reveal that semantic dimensions are structured according to distinct principles, with polarity emerging as a key factor driving their decomposition into subdimensions. The neural correlates of the identified subdimensions support their cognitive and neuroscientific plausibility.