Large Language Models

Overview

Hive AutoML offers training for two different types of large language models (LLMs): a general base LLM and chat LLM. The base LLM model can be used for a wide range of text generation tasks, while chat models are built specifically for interactions with the user (often in the format of asking questions and responding to questions). Chat models also have the added ability to incorporate conversation history into a prompt.

Which Base Model Should I Use?

Base Model	Latency	Max Tokens	Use Case	Dataset Validation
Hive LLM Chat - 70B	High	4096 per request	Higher latency and cost, stronger performance.	1. Minimum requirement of prompt and completion training data is 100 examples. 2. Each row of data (both prompt and completion added together) has a maximum length restriction of 16,384 characters.
Hive LLM Base - 70B	High	4096 per request	Higher latency and cost, suitable for non-chat applications.	1. Minimum requirement of prompt and completion training data is 100 examples. 2. Each row of data (both prompt and completion added together) has a maximum length restriction of 16,384 characters.