# Explainer: Latent Space Experts

DevFeed: [Explainer: Latent Space Experts](<https://devfeed.tech/articles/explainer-latent-space-experts-33449.md>)

Original publisher: [Read original article](<https://timkellogg.me/blog/2024/12/24/latent-experts>)

Published: 2024-12-24T00:00:00Z

Content type: opinion

Language: en

Sources: [Tim Kellogg](<https://devfeed.tech/sources/tim-kellogg.md>)

Topics: [Large Language Model](<https://devfeed.tech/topics/llm.md>), [Embeddings](<https://devfeed.tech/topics/embeddings.md>), [Google](<https://devfeed.tech/topics/google.md>), [Retrieval Augmented Generation (RAG)](<https://devfeed.tech/topics/retrieval-augmented-generation-rag.md>), [Cache](<https://devfeed.tech/topics/cache.md>)

Tags: [embeddings](<https://devfeed.tech/tags/embeddings.md>), [explainer](<https://devfeed.tech/tags/explainer.md>), [google](<https://devfeed.tech/tags/google.md>), [llms](<https://devfeed.tech/tags/llms.md>), [paper](<https://devfeed.tech/tags/paper.md>), [rag](<https://devfeed.tech/tags/rag.md>), [reasoning](<https://devfeed.tech/tags/reasoning.md>)

## AI overview

This explainer discusses Google DeepMind's paper "Deliberation in Latent Space via Differentiable Cache Augmentation." It describes pairing a generalist frozen LLM with a domain-specific coprocessor LLM that supplies additional embeddings, and compares this approach with retrieval-augmented generation and text-based communication.

## Source excerpt

A new paper just dropped from Google DeepMind, Deliberation in Latent Space via Differentiable Cache Augmentation. I don't think this paper is very readable, but it also seems quite important so I wanted to take a moment to break it down, as I understand it.