Mahal Systems HR-MultiWOZ
Fits HR help-desk chatbots that need conversation test cases. Borrow the LLM-generated dialogue pipeline built to avoid real HR conversations.
What it asks the agent to do
- hold a task-oriented HR conversation (benefits, leave, payroll questions)
- track dialogue state and slots across turns
- route an HR request to the right domain
What to borrow from it
privacy patterncase template
Caveats
A dialogue dataset, not an agent benchmark; no tools or state; no code link found.
Details
| Workloads | hr, customer support |
|---|---|
| Interaction | multi turn scripted |
| Environment | static |
| End state checked | No |
| Tools | none; dialogue dataset |
| Grading | exact match, human |
| Size | 550 |
| Data | synthetic |
| Personal data | none |
| License | CC-BY-4.0 |
| Harness | none |
| Status | frozen |