Clustering

As explained in How Large Language Models Work

A technique that groups data items according to similarities among their features or representations.

Clustering groups texts or other data items by similarity. With embeddings, each item becomes a numeric vector, so similar meanings land near one another even when the texts use different words. section 8.3.1

A clustering workflow typically does the following:

• Represent each text as an embedding produced by an embedding model. section 8.3.1
• Group nearby or otherwise similar embeddings into useful categories, using algorithms such as k-means or HDBSCAN. section 8.3.1
• Use the groups to inspect patterns, route related cases, or support decisions rather than assuming the LLM must solve the whole problem. section 8.3.1

Support requests while customers wait

A customer describes a problem, an embedding model represents it, and clustering groups the request with similar problems in the queue. Analysts can route those customers together, share a vetted solution for common cases, and let customers talk to a person when automation fails. section 8.3.2

Figure 8.4 LLMs produce numeric vectors known as embeddings as an intrinsic part of their function-ing. The utility of these embeddings is dependent on the fact that these numbers only change a little bit when given similar text. The two example tests here will have similar embeddings, and thus, their plots look similar, even though they don’t share any of the same words. This is a powerful feature that was present in older machine learning techniques.

Use the figure to see why clustering can group texts that share meaning without sharing exact words: similar texts produce similar vectors and nearby plotted positions. section 8.3.1

How How Large Language Models Work covers it

How deep it goes

Explains embeddings, grouping, named algorithms, and a support workflow, but does not provide algorithmic mechanics or parameter guidance.

Examples

Uses market-segment analysis, plotted embedding texts, and grouped customer problems as concrete examples.

What it assumes

Assumes familiarity with text embeddings at a basic level and introduces their use as numeric inputs for clustering.

What you can do after

Apply clustering to embedding-based workflows that organize similar cases and route them toward analysts or vetted solutions.

How it connects

Connects clustering with embeddings, outlier detection, information visualization, classification, regression, and customer support.

liveBookmanning.com

Read How Large Language Models Work on liveBook Get the book at manning.com

Also explained in

• 

  liveBookmanning.com
  Data Without LabelsClustering is an unsupervised technique that groups data points so that members of the same group are more similar to one another than to members of other groups.
• 

  liveBookmanning.com
  The Well-Grounded Data AnalystClustering is the process of grouping observations so that members of the same group are more similar to one another than to members of other groups.
• 

  liveBookmanning.com
  Outlier Detection in PythonAn unsupervised learning technique that divides items into groups whose members are more similar to one another than to members of other groups.
• 

  liveBookmanning.com
  Advanced Algorithms and Data StructuresA clustering algorithm groups unlabeled data points into subsets according to relationships or similarities among them.
• 

  liveBookmanning.com
  Grokking Machine LearningAn algorithm that partitions data points into groups based on similarity.
• 

  liveBookmanning.com
  The Art of AI Product DevelopmentClustering is an unsupervised learning technique that groups data points according to similarities in their characteristics.
• 

  liveBookmanning.com
  Statistics Slam DunkCustomer segmentation is the division of customers into groups based on shared purchasing, demographic, or other attributes.
• 

  liveBookmanning.com
  The AI Pocket BookAn unsupervised learning method that groups similar data points into clusters.
• 

  liveBookmanning.com
  First-Party Data ActivationThe practice of dividing a customer or user population into groups with shared characteristics or behaviors for differentiated marketing.
• 

  liveBookmanning.com
  Optimization Algorithms: AI techniques for design, planning, and control problemsThe task of partitioning objects into groups whose members are more similar to one another than to members of other groups.
• 

  liveBookmanning.com
  Classic Computer Science Problems in Swift: Essential techniques for practicing programmersClustering is a computational technique that divides data points into groups whose members are related according to some measure or interpretation.
• 

  liveBookmanning.com
  AI-Powered SearchClustering is an unsupervised learning technique that groups items according to similarity.
• 

  liveBookmanning.com
  Data Analysis with LLMsText clustering is the unsupervised grouping of text documents according to the similarity of their represented features.
• 

  liveBookmanning.com
  Inside AIClustering is the computational task of grouping data items according to similarities without requiring predefined categories.
• 

  liveBookmanning.com
  Data-Driven CybersecurityClustering is a data-analysis technique that groups observations according to similarity without requiring predefined labels.
• 

  liveBookmanning.com
  Learn Docker in a Month of Lunches, Second EditionA cluster is a group of machines managed together as a single computing system.
• 

  liveBookmanning.com
  Financial AI in PracticeClustering is the grouping of observations according to similarities in their measured attributes.

Compare how 18 books explain clustering →

Related concepts

• K-means clusteringproduced by · K-means clustering partitions data into a specified number of groups by assigning observations to representative centers and iteratively updating those centers to reduce within-group distances.
• k-means++ initializationincludes · K-means++ initialization is a procedure that selects initial centroids using probability proportional to squared distance from previously selected centroids.
• unsupervised learningkind of · Unsupervised learning is a machine-learning approach that discovers patterns, structure, or representations in data without labeled target outputs.
• DBSCANincludes · DBSCAN is a density-based clustering algorithm that groups observations in dense regions and identifies observations in lower-density regions as noise or outliers.
• hierarchical clusteringincludes · Hierarchical clustering is a clustering method that organizes observations into nested groups by successively merging or splitting clusters.
• embeddingused by · A numerical vector representation of an item or input that encodes features or relationships useful for computational tasks.

This explainer was generated by Manning's knowledge graph from the text of How Large Language Models Work. The section links point to the parts of the book it draws on.