World Info Activation Trace
World Info Activation Trace is a debugging feature introduced by Luker that helps you understand why a World Info entry was activated (or not activated).
Problem It Solves
When World Info configurations become complex (many entries, multiple keyword layers, different scan depths), users often encounter these confusions:
- Why wasn't a certain entry activated?
- Which keyword triggered this entry?
- In which message was the keyword matched?
- Is the scan depth setting reasonable?
Activation trace records the complete process from matching to injection for each entry, making these questions clear at a glance.
Trace Information
For each activated World Info entry, the trace record includes:
- Trigger keyword: Which keyword (or keyword combination) triggered the activation
- Match location: In which message the keyword was matched
- Scan depth: The current scan depth setting
- Activation method: Keyword match, constant (always active), probability activation, etc.
- Injection position: Where the entry was ultimately injected in the prompt
Usage
In the World Info editor, each entry has a fa-route icon button at its top-right; clicking it shows the most recent scan's activation trace for that entry:
The popup top shows the entry's scope, last activation time, and reason; below it, recent scans are listed in reverse chronological order — activated scans are marked red, non-activated ones grey. This makes it immediately clear which keyword in which message triggered the entry, and whether the scan depth is sufficient — so you can refine the keywords or scan depth accordingly.
Related Pages
- World Info Basics — Basic concepts of World Info
- Memory Graph — Memory Graph's World Info projection feature