# Extracting Conceptual Spaces from LLMs Using Prototype Embeddings

arXiv: https://arxiv.org/abs/2509.19269 · HTML: https://mlsanity.com/paper/2509.19269/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 27 of its 69 references that are in the map.
- First posted on arXiv 23 Sep 2025.

## Builds on (most cited first, 27 of 27)

- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks](https://mlsanity.com/paper/1908.10084/) (2019)
- [SimCSE: Simple Contrastive Learning of Sentence Embeddings](https://mlsanity.com/paper/2104.08821/) (2021)
- [How Contextual are Contextualized Word Representations? Comparing the Geometry of BERT, ELMo, and GPT-2 Embeddings](https://mlsanity.com/paper/1909.00512/) (2019)
- [On the Cross-lingual Transferability of Monolingual Representations](https://mlsanity.com/paper/1910.11856/) (2019)
- [NV-Embed: Improved Techniques for Training LLMs as Generalist Embedding Models](https://mlsanity.com/paper/2405.17428/) (2024)
- [LLM2Vec: Large Language Models Are Secretly Powerful Text Encoders](https://mlsanity.com/paper/2404.05961/) (2024)
- [Exploiting Similarities among Languages for Machine Translation](https://mlsanity.com/paper/1309.4168/) (2013)
- [Probing Pretrained Language Models for Lexical Semantics](https://mlsanity.com/paper/2010.05731/) (2020)
- [A robust self-learning method for fully unsupervised cross-lingual mappings of word embeddings](https://mlsanity.com/paper/1805.06297/) (2018)
- [Intermediate-Task Transfer Learning with Pretrained Models for Natural Language Understanding: When and Why Does It Work?](https://mlsanity.com/paper/2005.00628/) (2020)
- [Can Language Models Encode Perceptual Structure Without Grounding? A Case Study in Color](https://mlsanity.com/paper/2109.06129/) (2021)
- [Fast, Effective, and Self-Supervised: Transforming Masked Language Models into Universal Lexical and Sentence Encoders](https://mlsanity.com/paper/2104.08027/) (2021)
- [Large language models predict human sensory judgments across six modalities](https://mlsanity.com/paper/2302.01308/) (2023)
- [Things not Written in Text: Exploring Spatial Commonsense from Visual Signals](https://mlsanity.com/paper/2203.08075/) (2022)
- [The World of an Octopus: How Reporting Bias Influences a Language Model's Perception of Color](https://mlsanity.com/paper/2110.08182/) (2021)
- [Can Language Models Understand Physical Concepts?](https://mlsanity.com/paper/2305.14057/) (2023)
- [MirrorWiC: On Eliciting Word-in-Context Representations from Pretrained Language Models](https://mlsanity.com/paper/2109.09237/) (2021)
- [BERT Knows Punta Cana is not just beautiful, it's gorgeous: Ranking Scalar Adjectives with Contextualised Representations](https://mlsanity.com/paper/2010.02686/) (2020)
- [Modelling Commonsense Properties using Pre-Trained Bi-Encoders](https://mlsanity.com/paper/2210.02771/) (2022)
- [Distilling Semantic Concept Embeddings from Contrastively Fine-Tuned Language Models](https://mlsanity.com/paper/2305.09785/) (2023)
- [Sniff AI: Is My 'Spicy' Your 'Spicy'? Exploring LLM's Perceptual Alignment with Human Smell Experiences](https://mlsanity.com/paper/2411.06950/) (2024)
- [Do ever larger octopi still amplify reporting biases? Evidence from judgments of typical colour](https://mlsanity.com/paper/2209.12786/) (2022)
- [Adjusting Interpretable Dimensions in Embedding Space with Human Judgments](https://mlsanity.com/paper/2404.02619/) (2024)
- [Cabbage Sweeter than Cake? Analysing the Potential of Large Language Models for Learning Conceptual Spaces](https://mlsanity.com/paper/2310.05481/) (2023)
- [TouchAI: Exploring human-AI perceptual alignment in touch through language model representations](https://mlsanity.com/paper/2406.06587/) (2024)
- [Concept Layers: Enhancing Interpretability and Intervenability via LLM Conceptualization](https://mlsanity.com/paper/2502.13632/) (2025)

## Abstract (arXiv)

Conceptual spaces represent entities and concepts using cognitively meaningful dimensions, typically referring to perceptual features. Such representations are widely used in cognitive science and have the potential to serve as a cornerstone for explainable AI. Unfortunately, they have proven notoriously difficult to learn, although recent LLMs appear to capture the required perceptual features to a remarkable extent. Nonetheless, practical methods for extracting the corresponding conceptual spaces are currently still lacking. While various methods exist for extracting embeddings from LLMs, extracting conceptual spaces also requires us to encode the underlying features. In this paper, we propose a strategy in which features (e.g. sweetness) are encoded by embedding the description of a corresponding prototype (e.g. a very sweet food). To improve this strategy, we fine-tune the LLM to align the prototype embeddings with the corresponding conceptual space dimensions. Our empirical analysis finds this approach to be highly effective.