General NLU Performance on NLU Average
0.874Average AccuracyDeBERTa-V3 base + PiFi
Evaluation results
| Method | Links | ||
|---|---|---|---|
| DeBERTa-V3 base + PiFiModel Backbone=DeBERTa-V3-base, PiFi Integration=true, Source LLM=Llama-3.1-8B2025.06 | 0.874 | 0.8287 | |
| ELECTRA base + PiFiModel Backbone=ELECTRA-base, PiFi Integration=true, Source LLM=Llama-3.1-8B2025.06 | 0.8671 | 0.8076 | |
| DeBERTa-V3 baseModel Backbone=DeBERTa-V3-base2025.06 | 0.8634 | 0.8162 | |
| DeBERTa base + PiFiModel Backbone=DeBERTa-base, PiFi Integration=true, Source LLM=Llama-3.1-8B2025.06 | 0.8563 | 0.8037 | |
| RoBERTa base + PiFiModel Backbone=RoBERTa-base, PiFi Integration=true, Source LLM=Llama-3.1-8B2025.06 | 0.8542 | 0.798 | |
| BERT base + PiFiModel Backbone=BERT-base, PiFi Integration=true, Source LLM=Llama-3.1-8B2025.06 | 0.8475 | 0.7953 | |
| DeBERTa baseModel Backbone=DeBERTa-base2025.06 | 0.844 | 0.7943 | |
| RoBERTa baseModel Backbone=RoBERTa-base2025.06 | 0.8432 | 0.7879 | |
| BERT baseModel Backbone=BERT-base2025.06 | 0.8245 | 0.7752 | |
| ELECTRA baseModel Backbone=ELECTRA-base2025.06 | 0.82 | 0.7662 |