Skip to content

Latest commit

 

History

History
93 lines (92 loc) · 3.53 KB

chatml.md

File metadata and controls

93 lines (92 loc) · 3.53 KB

(This document is a preview of the underlying format consumed by ChatGPT models. As a developer, you can use our higher-level API and won't need to interact directly with this format today — but expect to have the option in the future!)

Traditionally, GPT models consumed unstructured text. ChatGPT models instead expect a structured format, called Chat Markup Language (ChatML for short). ChatML documents consist of a sequence of messages. Each message contains a header (which today consists of who said it, but in the future will contain other metadata) and contents (which today is a text payload, but in the future will contain other datatypes). We are still evolving ChatML, but the current version (ChatML v0) can be represented with our upcoming "list of dicts" JSON format as follows:

[
 {"token": "<|im_start|>"},
 "system\nYou are ChatGPT, a large language model trained by OpenAI. Answer as concisely as possible.\nKnowledge cutoff: 2021-09-01\nCurrent date: 2023-03-01",
 {"token": "<|im_end|>"}, "\n", {"token": "<|im_start|>"},
 "user\nHow are you",
 {"token": "<|im_end|>"}, "\n", {"token": "<|im_start|>"},
 "assistant\nI am doing well!",
 {"token": "<|im_end|>"}, "\n", {"token": "<|im_start|>"},
 "user\nHow are you now?",
 {"token": "<|im_end|>"}, "\n"
]

You could also represent it in the classic "unsafe raw string" format. However, this format inherently allows injections from user input containing special-token syntax, similar to SQL injections:

<|im_start|>system
You are ChatGPT, a large language model trained by OpenAI. Answer as concisely as possible.
Knowledge cutoff: 2021-09-01
Current date: 2023-03-01<|im_end|>
<|im_start|>user
How are you<|im_end|>
<|im_start|>assistant
I am doing well!<|im_end|>
<|im_start|>user
How are you now?<|im_end|>

Non-chat use-cases

ChatML can be applied to classic GPT use-cases that are not traditionally thought of as chat. For example, instruction following (where a user requests for the AI to complete an instruction) can be implemented as a ChatML query like the following:

[
 {"token": "<|im_start|>"},
 "user\nList off some good ideas:",
 {"token": "<|im_end|>"}, "\n", {"token": "<|im_start|>"},
 "assistant"
]

We do not currently allow autocompleting of partial messages,

[
 {"token": "<|im_start|>"},
 "system\nPlease autocomplete the user's message.",
 {"token": "<|im_end|>"}, "\n", {"token": "<|im_start|>"},
 "user\nThis morning I decided to eat a giant"
]

Note that ChatML makes explicit to the model the source of each piece of text, and particularly shows the boundary between human and AI text. This gives an opportunity to mitigate and eventually solve injections, as the model can tell which instructions come from the developer, the user, or its own input.

Few-shot prompting

In general, we recommend adding few-shot examples using separate system messages with a name field of example_user or example_assistant. For example, here is a 1-shot prompt:

<|im_start|>system
Translate from English to French
<|im_end|>
<|im_start|>system name=example_user
How are you?
<|im_end|>
<|im_start|>system name=example_assistant
Comment allez-vous?
<|im_end|>
<|im_start|>user
{{user input here}}<|im_end|>

If adding instructions in the system message doesn't work, you can also try putting them into a user message. (In the near future, we will train our models to be much more steerable via the system message. But to date, we have trained only on a few system messages, so the models pay much more attention to user examples.)