Feedback Operator
The Feedback LOP manages conversation state by copying data from a referenced table into its internal conversation DAT. It provides a simple reset mechanism for returning conversation state to a known starting point, making it useful for establishing example conversations, enforcing patterns for LLM agents, or resetting mid-workflow.
Overview
Section titled “Overview”The Feedback operator holds a conversation table internally. During normal operation, it copies data from its internal input table into the conversation DAT. When reset is triggered, it copies the contents of its first internal input back to the conversation DAT, restoring the original state.
This is one of the simplest LOPs operators, inheriting directly from DotChatUtil rather than DotLOPUtils. It has no model page, no API calls, and no tool integration.
Input/Output
Section titled “Input/Output”Inputs
Section titled “Inputs”None wired. The operator references an external table through the Target DAT parameter on the Feedback page.
Outputs
Section titled “Outputs”- Output 1 (Table DAT): The current conversation state with columns:
role,message,id,timestamp.
Usage Examples
Section titled “Usage Examples”Setting Up a Conversation Reset Point
Section titled “Setting Up a Conversation Reset Point”- Create a table DAT containing your starting conversation (columns:
role,message,id,timestamp). - Place a Feedback LOP in your network.
- On the Feedback page, set Target DAT to reference your starting conversation table.
- The output will reflect the current conversation state.
- Pulse Resetpulse at any time to restore the conversation back to the initial state from the referenced table.
Using Reset Toggle for Persistent Override
Section titled “Using Reset Toggle for Persistent Override”- Toggle Reset to On on the Feedback page.
- While Reset is active, the operator’s color changes to indicate its state.
- Toggle Reset back to Off to resume normal operation.
Common Use Cases
Section titled “Common Use Cases”- Establishing a starting point for LLM agent conversations with pre-filled example exchanges.
- Enforcing specific conversation patterns or structures that agents should follow.
- Resetting conversation state to a predefined baseline between test runs.
- Synchronizing context between multiple operators in a multi-agent workflow.
Performance Considerations
Section titled “Performance Considerations”This is a lightweight operator. Table copy operations are fast and suitable for real-time use. Resetting very large conversation tables frequently may cause minor performance impact.
Parameters
Section titled “Parameters”Feedback
Section titled “Feedback”op('feedback').par.Targetdat OP - Default:
"" (Empty String)
op('feedback').par.Reset Toggle - Default:
False
op('feedback').par.Resetpulse Pulse - Default:
False
Changelog
Section titled “Changelog”v1.0.02024-11-06
Initial release