General Knowledge Question Answering on Commonsense+TriviaQA
69.6AccuracyDUET-IF
Evaluation Results
| Method | Links | |
|---|---|---|
| DUET-IFBackbone=Llama-3-8b-Instruct2025.02 | 69.6 | |
| TriviaQATrain=TriviaQA, Backbone=Llama-3-8b-Instruct2025.02 | 66.5 | |
| CommonsenseQATrain=CommonsenseQA, Backbone=Llama-3-8b-Instruct2025.02 | 65.8 | |
| gsm8kTrain=gsm8k, Backbone=Llama-3-8b-Instruct2025.02 | 64.1 | |
| SciQTrain=SciQ, Backbone=Llama-3-8b-Instruct2025.02 | 63.4 | |
| TruthfulQATrain=TruthfulQA, Backbone=Llama-3-8b-Instruct2025.02 | 61 | |
| HellaswagTrain=Hellaswag, Backbone=Llama-3-8b-Instruct2025.02 | 60.4 | |
| HeadQATrain=HeadQA, Backbone=Llama-3-8b-Instruct2025.02 | 60 | |
| WikitextTrain=Wikitext, Backbone=Llama-3-8b-Instruct2025.02 | 59.9 | |
| PubmedTrain=Pubmed, Backbone=Llama-3-8b-Instruct2025.02 | 58.4 |