I've heard that conversation data in the ChatGPT Team Plan is not used for model training, but we would like to contribute our data to aid the model's development. Is there a setting within the Team Plan that allows us to provide our data?
#Data contribution
south anvil
mental doveBOT
Hi @south anvil. I'm kapa.ai, your friendly support bot. I will try to answer your question by searching through the knowledge base. :wave:
Based on the information provided in the extracts, OpenAI does not use content from business offerings such as the ChatGPT Team plan to train their models. This is part of their commitment to data privacy and security for businesses. As such, there is no setting within the Team Plan that allows you to provide your data for model training.
For more details, you can refer to the Enterprise Privacy page (https://openai.com/enterprise-privacy) and the Data Controls FAQ (https://help.openai.com/en/articles/7730893-data-controls-faq) mentioned in the extracts.
Note: you can ask me a follow up question by @ mentioning me again :speech_balloon:
south anvil
@mental dove
I have another question. I heard that the ChatGPT Team Plan supports up to 32k tokens. What is the token limit for the standard GPT-4?
mental doveBOT
Let me check the knowledge base for you. One moment! :books:
The token limit for the standard GPT-4 model varies depending on the specific version of the model. Here are the token limits for some versions of GPT-4 based on the provided extracts:
gpt-4-1106-vision-preview: 128,000 tokens for input, but can only output up to 4,096 tokens (source (https://platform.openai.com/docs/models))gpt-4-0613: 8,192 tokens (source (https://platform.openai.com/docs/models))gpt-4-32k-0613: 32,768 tokens (source (https://platform.openai.com/docs/models))gpt-4-0125-previewandgpt-4-turbo-preview: 128,000 tokens for input, but can only output up to 4,096 tokens (source (https://platform.openai.com/docs/models))
Please note that while the input token limit can be quite large for some models, the output token limit is often much smaller. For example, the gpt-4-1106-preview model can accept up to 128,000 input tokens but can only output up to 4,096 tokens (source (https://community.openai.com/t/515790)).
Also, remember that the rate limiter considers both input and output tokens. For example, at tier 1 (paying less than $50 in the past), gpt-4-turbo-preview has a limit of 150,000 tokens per minute and a more restrictive 500,000 tokens per day. The standard gpt-4 has a limit of 10,000 tokens per minute with no daily limit (source (https://community.openai.com/t/612242)).
For more information on token limits and how to work within them, you can refer to this help article (https://help.openai.com/en/articles/4936856-what-are-tokens-and-how-to-count-them#h_051eb08805).
Note: you can ask me a follow up question by @ mentioning me again :speech_balloon:
south anvil
@mental dove
Are the token count and context window in ChatGPT compatible? For example, does increasing the allowable number of input tokens also increase the context window?
mental doveBOT
Let me check the knowledge base for you. One moment! :books:
The token count and the context window in ChatGPT are related but not directly proportional. The context window refers to the maximum number of tokens that the model can consider from the conversation history when generating a response. On the other hand, the token count refers to the total number of tokens in both the input and output.
Increasing the allowable number of input tokens does not necessarily increase the context window. The context window is a fixed property of the model and depends on the specific version of the model you are using. For example, the gpt-3.5-turbo-1106 model has a context window of 16,385 tokens (source (https://community.openai.com/t/609532)), while the gpt-4 model has a context window of 8,000 tokens and the gpt-4-turbo has a context window of 125,000 tokens (source (https://community.openai.com/t/612242)).
However, if the total number of tokens in your conversation (including both input and output tokens) exceeds the context window, the model will start to drop the oldest messages to fit within the context window (source (https://platform.openai.com/docs/assistants/how-it-works)). This means that if you increase the number of input tokens, you may end up with fewer tokens available for the model's output.
In summary, while the token count and context window are related, they are not directly proportional, and increasing the number of input tokens does not increase the context window.
Note: you can ask me a follow up question by @ mentioning me again :speech_balloon:
south anvil
@mental dove
My understanding is that the Team Plan currently supports 32k tokens. Is that correct?
mental doveBOT
Let me check the knowledge base for you. One moment! :books:
Based on the provided extracts, there is no specific information about the token limit for the ChatGPT Team Plan. However, it's mentioned that different models have different token limits. For instance, the gpt-4-32k-0613 model has a token limit of 32,768 tokens (source (https://platform.openai.com/docs/models)).
For the most accurate information, you should refer to the specific details of the Team Plan or the documentation for the model you are using.
Note: you can ask me a follow up question by @ mentioning me again :speech_balloon:
south anvil
@mental dove
Thanks kapa!