Chat completion model

As explained in AI Agents in Action

A chat completion model is a language model designed to generate responses across iterative conversational exchanges.

A chat completion model generates responses through iterative conversational exchanges. Unlike completion models, it benefits from refinement, making it suited to reasoning, planning, agents, and assistants. section 2.0

ChatGPT and iterative agents

ChatGPT illustrates a chat completion model refined with reinforcement learning from human feedback. An agent can use the same conversational behavior to revisit a task, improve its response, and plan across multiple exchanges. section 2.0

Evaluate a model for an agent by checking:

• Performance on the target task, including coding, reasoning, or planning. section 2.4
• Model size, training input, training method, and context-token capacity. section 2.4
• Context capacity, especially when multiple agents converse over one task. section 2.4

Figure 2.11 The important criteria to consider when consuming an LLM

Use the criteria overview to compare models before consuming one in an agent application. section 2.4

How AI Agents in Action covers it

How deep it goes

Distinguishes chat completion from completion models and surveys selection criteria without detailing implementation internals.

Examples

Uses ChatGPT, iterative conversations, planning agents, and multi-agent context as examples.

What it assumes

Assumes basic awareness of language models while introducing the distinction between chat completion and completion models.

What you can do after

Evaluate models for agent applications by comparing task performance, size, training, and context capacity.

How it connects

Connects chat completion models with large language models, GPT, reinforcement learning from human feedback, completion models, and AI agents.

liveBookmanning.com

Read AI Agents in Action on liveBook Get the book at manning.com

Related concepts

• ChatOpenAIincludes · ChatOpenAI is a LangChain integration that provides an interface for invoking OpenAI chat-oriented language models.
• prompt engineeringused by · Prompt engineering is the deliberate design and refinement of instructions, context, examples, or other inputs to guide an AI model’s behavior or outputs.
• pipelineused by · A pipeline is an ordered or coordinated sequence of connected processing stages or tasks that passes data, work, or results toward an outcome.
• AI agentuses · An AI agent is a computational system that uses artificial intelligence to interpret information, select actions, and pursue goals through interaction with an environment.

This explainer was generated by Manning's knowledge graph from the text of AI Agents in Action. The section links point to the parts of the book it draws on.