Accessing this feature
Your access to the feature described in this article depends on your license package and pricing plan.
To learn which features are available to your organization and how to add more, contact your Hyperscience representative.
With the General Prompting Block, you can leverage the power of ORCA VLMs in custom flows for use cases that don't require leveraging layouts or releases.
The General Prompting Block can be an effective solution for testing prompts quickly and can power features like Document Chat in place of third-party VLMs.
Example uses
Incorporating ORCA VLMs in Document Chat in place of third-party LLMs/VLMs. This application of the General Prompting Block makes it possible to use Document Chat in environments that cannot be connected to the internet or for use cases where third-party LLMs/VLMs cannot be used for security reasons.
Creating complex tasks that span across pages or documents. In these cases, layouts can be too restrictive, and it may be more efficient to try several different kinds of prompts in succession without having to redeploy layouts.
Implementing use cases beyond extraction. We can use the General Prompting Block to classify documents, normalize and concatenate data, and even act as a second layer of validation before a human-in-the-loop (HiTL) task. There are many possibilities beyond extraction.
Limitations
The output of the General Prompting Block is available only in its Output tab in the block’s specific flow run. To learn more about flow runs, see Testing and Debugging Flows.
Prerequisites
The prerequisites outlined for ORCA VLMs in ORCA (Optical Reasoning and Cognition Agent) VLMs
A custom flow where the General Prompting Block is:
preceded by an Install LLM/ VLM Block and a Code Block and directly preceded by a Code Block. These blocks create and pass the input values required by the General Prompting Block.
directly followed by a Code Block. This block will take the output of the General Prompting Block and create the necessary input for the next block in the flow.
How it works
You can learn more about the inputs and outputs of the General Prompting Block in the Flows SDK documentation.