Saved in:
Bibliographic Details
Main Authors: Tennenholtz, Guy, Chow, Yinlam, Hsu, Chih-Wei, Jeong, Jihwan, Shani, Lior, Tulepbergenov, Azamat, Ramachandran, Deepak, Mladenov, Martin, Boutilier, Craig
Format: Preprint
Published: 2023
Subjects:
Online Access:https://arxiv.org/abs/2310.04475
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866929274450083840
author Tennenholtz, Guy
Chow, Yinlam
Hsu, Chih-Wei
Jeong, Jihwan
Shani, Lior
Tulepbergenov, Azamat
Ramachandran, Deepak
Mladenov, Martin
Boutilier, Craig
author_facet Tennenholtz, Guy
Chow, Yinlam
Hsu, Chih-Wei
Jeong, Jihwan
Shani, Lior
Tulepbergenov, Azamat
Ramachandran, Deepak
Mladenov, Martin
Boutilier, Craig
contents Embeddings have become a pivotal means to represent complex, multi-faceted information about entities, concepts, and relationships in a condensed and useful format. Nevertheless, they often preclude direct interpretation. While downstream tasks make use of these compressed representations, meaningful interpretation usually requires visualization using dimensionality reduction or specialized machine learning interpretability methods. This paper addresses the challenge of making such embeddings more interpretable and broadly useful, by employing Large Language Models (LLMs) to directly interact with embeddings -- transforming abstract vectors into understandable narratives. By injecting embeddings into LLMs, we enable querying and exploration of complex embedding data. We demonstrate our approach on a variety of diverse tasks, including: enhancing concept activation vectors (CAVs), communicating novel embedded entities, and decoding user preferences in recommender systems. Our work couples the immense information potential of embeddings with the interpretative power of LLMs.
format Preprint
id arxiv_https___arxiv_org_abs_2310_04475
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Demystifying Embedding Spaces using Large Language Models
Tennenholtz, Guy
Chow, Yinlam
Hsu, Chih-Wei
Jeong, Jihwan
Shani, Lior
Tulepbergenov, Azamat
Ramachandran, Deepak
Mladenov, Martin
Boutilier, Craig
Computation and Language
Artificial Intelligence
Machine Learning
Embeddings have become a pivotal means to represent complex, multi-faceted information about entities, concepts, and relationships in a condensed and useful format. Nevertheless, they often preclude direct interpretation. While downstream tasks make use of these compressed representations, meaningful interpretation usually requires visualization using dimensionality reduction or specialized machine learning interpretability methods. This paper addresses the challenge of making such embeddings more interpretable and broadly useful, by employing Large Language Models (LLMs) to directly interact with embeddings -- transforming abstract vectors into understandable narratives. By injecting embeddings into LLMs, we enable querying and exploration of complex embedding data. We demonstrate our approach on a variety of diverse tasks, including: enhancing concept activation vectors (CAVs), communicating novel embedded entities, and decoding user preferences in recommender systems. Our work couples the immense information potential of embeddings with the interpretative power of LLMs.
title Demystifying Embedding Spaces using Large Language Models
topic Computation and Language
Artificial Intelligence
Machine Learning
url https://arxiv.org/abs/2310.04475