DSTC12 Track 2 Submission Form
Please note: this form is limited to **1 submission per team**, make sure it's your final one!
Team name
Team leader's name
Team leader's email
Team leader's affiliation
Other team members' names and affiliations - 1 record per line of the following format: name, affiliation
What LLM type did you use?
Open-source - self-hosted
Proprietary via API
No LLM
Other
How large of an LLM did you use?
<30B
30-100B
>100B
Unknown (proprietary API)
No LLM
Other
Did you use any conversational information (previous / past context of the utterance)? Please specify if yes
What clustering algorithm did you use?
What text embedding model did you use?
Did you use an embedding dimensionality reduction technique? (Please specify which one if yes)
Did you use a data augmentation technique (please specify what kind)?
How did you use the should-link / cannot-link pairs?
Did you use the theme label style guide - if yes, how?
Short (1-2 paragraph) description of your approach
Inferenced AppenTravel testset in a zip file (all-to-predict.jsonl with 'theme_label_predicted' field)
Browse Files
Drag and drop files here
Choose a file
Cancel
of
Save
Submit
Should be Empty: