# Latent Topology Induction for Understanding Contextualized Representations

arXiv: https://arxiv.org/abs/2206.01512 · HTML: https://mlsanity.com/paper/2206.01512/

## Key facts

- Cited by 7 papers in the map (data through 14 Oct 2026).
- Builds on 24 of its 48 references that are in the map.
- First posted on arXiv 3 Jun 2022.

## Builds on (most cited first, 24 of 24)

- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [Auto-Encoding Variational Bayes](https://mlsanity.com/paper/1312.6114/) (2013)
- [BERT Rediscovers the Classical NLP Pipeline](https://mlsanity.com/paper/1905.05950/) (2019)
- [A Primer in BERTology: What we know about how BERT works](https://mlsanity.com/paper/2002.12327/) (2020)
- [Designing and Interpreting Probes with Control Tasks](https://mlsanity.com/paper/1909.03368/) (2019)
- [What do you learn from context? Probing for sentence structure in contextualized word representations](https://mlsanity.com/paper/1905.06316/) (2019)
- [Linguistic Knowledge and Transferability of Contextual Representations](https://mlsanity.com/paper/1903.08855/) (2019)
- [Visualizing and Measuring the Geometry of BERT](https://mlsanity.com/paper/1906.02715/) (2019)
- [Finding Universal Grammatical Relations in Multilingual BERT](https://mlsanity.com/paper/2005.04511/) (2020)
- [Perturbed Masking: Parameter-free Probing for Analyzing and Interpreting BERT](https://mlsanity.com/paper/2004.14786/) (2020)
- [Conditional probing: measuring usable information beyond a baseline](https://mlsanity.com/paper/2109.09234/) (2021)
- [What Do Neural Networks Learn When Trained With Random Labels?](https://mlsanity.com/paper/2006.10455/) (2020)
- [Differentiable Dynamic Programming for Structured Prediction and Attention](https://mlsanity.com/paper/1802.03676/) (2018)
- [Discovering Latent Concepts Learned in BERT](https://mlsanity.com/paper/2205.07237/) (2022)
- [Probing BERT in Hyperbolic Spaces](https://mlsanity.com/paper/2104.03869/) (2021)
- [Are Pre-trained Language Models Aware of Phrases? Simple but Strong Baselines for Grammar Induction](https://mlsanity.com/paper/2002.00737/) (2020)
- [A Tale of a Probe and a Parser](https://mlsanity.com/paper/2005.01641/) (2020)
- [Emergence of Separable Manifolds in Deep Language Representations](https://mlsanity.com/paper/2006.01095/) (2020)
- [Asking without Telling: Exploring Latent Ontologies in Contextual Representations](https://mlsanity.com/paper/2004.14513/) (2020)
- [Obtaining Better Static Word Embeddings Using Contextual Embedding Models](https://mlsanity.com/paper/2106.04302/) (2021)
- [Posterior Control of Blackbox Generation](https://mlsanity.com/paper/2005.04560/) (2020)
- [Latent Template Induction with Gumbel-CRFs](https://mlsanity.com/paper/2011.14244/) (2020)
- [Conditional Random Field Autoencoders for Unsupervised Structured Prediction](https://mlsanity.com/paper/1411.1147/) (2014)
- [Scaling Structured Inference with Randomization](https://mlsanity.com/paper/2112.03638/) (2021)

## Built on by (most cited first, 7 of 7)

- [M-FLAG: Medical Vision-Language Pre-training with Frozen Language Models and Latent Space Geometry Optimization](https://mlsanity.com/paper/2307.08347/) (2023)
- [On the Transformation of Latent Space in Fine-Tuned NLP Models](https://mlsanity.com/paper/2210.12696/) (2022)
- [Can LLMs facilitate interpretation of pre-trained language models?](https://mlsanity.com/paper/2305.13386/) (2023)
- [Exploring Alignment in Shared Cross-lingual Spaces](https://mlsanity.com/paper/2405.14535/) (2024)
- [Scaling up Discovery of Latent Concepts in Deep NLP Models](https://mlsanity.com/paper/2308.10263/) (2023)
- [NxPlain: Web-based Tool for Discovery of Latent Concepts](https://mlsanity.com/paper/2303.03019/) (2023)
- [CuMA: Aligning LLMs with Sparse Cultural Values via Demographic-Aware Mixture of Adapters](https://mlsanity.com/paper/2601.04885/) (2026)

## Abstract (arXiv)

In this work, we study the representation space of contextualized embeddings and gain insight into the hidden topology of large language models. We show there exists a network of latent states that summarize linguistic properties of contextualized representations. Instead of seeking alignments to existing well-defined annotations, we infer this latent network in a fully unsupervised way using a structured variational autoencoder. The induced states not only serve as anchors that mark the topology (neighbors and connectivity) of the representation manifold but also reveal the internal mechanism of encoding sentences. With the induced network, we: (1). decompose the representation space into a spectrum of latent states which encode fine-grained word meanings with lexical, morphological, syntactic and semantic information; (2). show state-state transitions encode rich phrase constructions and serve as the backbones of the latent space. Putting the two together, we show that sentences are represented as a traversal over the latent network where state-state transition chains encode syntactic templates and state-word emissions fill in the content. We demonstrate these insights with extensive experiments and visualizations.