Edge 293: Instruction Following Language Models

Instruction following LLs, OpenAI's InstructGPT and the Dust LLM framework.

Created Using Midjourney

In this Issue:

1. The Concept: An overview of instruction following language models.

2. The Research: A review of OpenAI’s InstructGPT paper.

3. The Tech: Some details about the Dust LLM framework.

💡 ML Concept of the Day: Instruction Following Language Models   

In a previous installment of this series, we discussed some of the ideas behind reinforcement learning with human feedback(RLHF) as a core component of modern large language models(LLMs). Arguably, the most relevant case of RLHF is to train LLMs to follow instructions. The main research in this area was unveiled by OpenAI with a model called InstructGPT.