Search Relevance on ESCI
61.03Macro F1DeBERTa-v3-large
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| DeBERTa-v3-largeFramework=CE, w/ fts=true2025.11 | 61.03 | 75.92 | |
| DeBERTa-v3-largeFramework=CE, w/ fts=false2025.11 | 59.01 | 75.37 | |
| DeBERTa-v3-base + BigBird-base + CoCoLM-baseFramework=Ensemble, w/ fts=true2025.11 | 58.27 | 75.65 | |
| BERT-largeFramework=CE, w/ fts=true2025.11 | 57.61 | 74.01 | |
| DeBERTa-v3-base + BigBird-base + CoCoLM-baseFramework=Ensemble, w/ fts=false2025.11 | 56.96 | 75.26 | |
| BERT-largeFramework=CE, w/ fts=false2025.11 | 55.82 | 73.36 | |
| BERT-largeFramework=BE, w/ fts=true2025.11 | 49.71 | 68.57 | |
| BERT-largeFramework=BE, w/ fts=false2025.11 | 48.98 | 69.77 | |
| DeBERTa-v3-largeFramework=BE, w/ fts=true2025.11 | 48.7 | 68.27 | |
| DeBERTa-v3-largeFramework=BE, w/ fts=false2025.11 | 46.87 | 67.25 | |
| gte-Qwen2-7BFramework=BE, w/ fts=true2025.11 | 44.66 | 68.37 | |
| gte-Qwen2-7BFramework=BE, w/ fts=false2025.11 | 42.95 | 67.81 | |
| Llama-3.1-8BFramework=LLM Inference, w/ fts=true2025.11 | 36.26 | 61.68 | |
| Llama-3.1-8BFramework=LLM Inference, w/ fts=false2025.11 | 35.25 | 59.83 |