End-to-End Task-Oriented Dialogue on MultiWOZ 2.2 (test)
0.379JGAFNCTOD-LLAMA2-13B
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| FNCTOD-LLAMA2-13BProtocol=Fine-tuned, Backbone=LLAMA2-13B2024.02 | 0.379 | 0.444 | |
| VICUNA-13B-v1.5Protocol=Zero-shot prompting, Approach=FNCTOD, Backbone=VICUNA-13B-v1.52024.02 | 0.338 | 0.231 | |
| BAICHUAN2-13B-CHATProtocol=Zero-shot prompting, Approach=FNCTOD, Backbone=BAICHUAN2-13B-CHAT2024.02 | 0.33 | 0.457 | |
| ZEPHYR-7B-BETAProtocol=Zero-shot prompting, Approach=FNCTOD, Backbone=ZEPHYR-7B-BETA2024.02 | 0.323 | 0.575 | |
| VICUNA-7B-v1.5Protocol=Zero-shot prompting, Approach=FNCTOD, Backbone=VICUNA-7B-v1.52024.02 | 0.294 | 0.377 | |
| ChatGPTProtocol=Zero-shot prompting2024.02 | 0.27 | 0.44 | |
| LLAMA2-13B-CHATProtocol=Zero-shot prompting, Approach=FNCTOD, Backbone=LLAMA2-13B-CHAT2024.02 | 0.258 | 0.277 | |
| LLAMA2-7B-CHATProtocol=Zero-shot prompting, Approach=FNCTOD, Backbone=LLAMA2-7B-CHAT2024.02 | 0.167 | 0.249 |