# OpenAI — Aligning language models to follow instructions

- Company: OpenAI (openai.com)
- Announced: 2022-01-27T08:00:00+00:00
- Category: new-model
- Coverage: not counted
- Announcement: yes
- Group: models
- Source: https://openai.com/index/instruction-following
- Record: https://forck.live/items/918-aligning-language-models-to-follow-instructions
- Subject: GPT / ChatGPT / API
- Models affected: InstructGPT

OpenAI has trained InstructGPT models that are better at following instructions than GPT-3, and they are now the default models on the API.

## Evidence

Verbatim from https://openai.com/index/instruction-following:

> We’ve trained language models that are much better at following user intentions than GPT-3 while also making them more truthful and less toxic, using techniques developed through our alignment research. These InstructGPT models, which are trained with humans in the loop, are now deployed as the default language models on our API.

---

Record: https://forck.live/items/918-aligning-language-models-to-follow-instructions
Catalogue: https://forck.live/llms.txt
Current issue: https://forck.live/feed.md