PaLM 2 Chat Bison

The chat-bison model is a large language model that excels at language understanding, language generation and conversations. This chat model is fine-tuned to conduct natural multi-turn conversations.

The PaLM 2 Chat Bison is ideal for text tasks that require back-and-forth interactions. For text tasks that can be completed with one API response (without the need for continuous conversation), use the PaLM 2 Text Bison.

Model capabilities

• Text in, text out.
• Generates text in a conversational format.
• Optimized for natural language tasks like chat bots or help agents.
• Can handle zero, one, and few-shot tasks (you provide no examples, one example, or several examples to inform model inference).

Model properties

• Training data: up to Feb 2023
• Max turns : 2,500

Model safety

• No adjustable safety settings.
• Error message instead of generated response for filtered input/output.
• No safety rating or score returned with response.

• Customer Service: Instruct the model to respond as customer service agents that only talk about your company's product
• Technical Support: Instruct the model to interact with customers as a call center agent with specific parameters about how to respond and what not to say
• Personas and characters: Instruct the model to respond in the style of a specific person ("...in the style of Shakespeare")
• Website companion: Create a conversational assistant for shopping, travel and other use cases

Intended use

• Your use of PaLM 2 is subject to the Generative AI Prohibited Use Policy.

Get started

You can use Vertex AI Studio to experiment with the chat-bison model in the Google Cloud console. The chat-bison model also supports Python.

Setup the PaLM 2 Chat Bison

To use the PaLM 2 Chat Bison in the Google Cloud console, open it in Vertex AI Studio. To use it programmatically, you need to set it up in Python. Use the following topics to get started:

Open the PaLM 2 Chat Bison in Vertex AI Studio (console)

You can try a demo chatbot based on chat-bison@001 in Vertex AI Studio by performing the following steps:

1. In the Vertex AI Studio, click Chat prompt.

2. In the Context textbox, enter the following context:

3. In the text input field, ask questions about astronomy.

Setup the PaLM 2 Chat Bison (Python)

To use the chat-bison model with Python, do the following:

Setup

1. Install Python 3.7 or higher if it's not already installed.
2. Clone the repository with Google Cloud Python samples:

3. Obtain authentication credentials. Create local credentials by running the following command and following the oauth2 flow (read more about the command here:

   Read more about Google Cloud Platform Authentication.

How to run a sample

1. Run the following code to navigate to the folder with generative AI samples:

2. Create a virtualenv. Samples are compatible with Python 3.7+.

3. Install the dependencies needed to run the samples.

4. The test file chat.py contains sample code that uses the chat-bison@001 model. The sample code asks the model how many planets there are in the solar system and contains context and an example. To run the sample, run the following command:

Versions

Links

• Test chat prompts