Named Entity Recognition on Dynamic-NER
43.4Macro F1Gemma3-27B
Evaluation results
| Method | Links | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Gemma3-27B2026.01 | 43.4 | — | — | — | — | — | — | — | — | |
| OTTERArchitecture=Cross-Encoder, Backbone=mmBERT, Training Steps=100k, Thresholding Strategy=Best per language2026.01 | 38.5 | — | — | — | — | — | — | — | — | |
| OTTERArchitecture=Cross-Encoder, Backbone=mmBERT, Training Steps=100k2026.01 | 38.2 | — | — | — | — | — | — | — | — | |
| Qwen3-32B2026.01 | 36.5 | — | — | — | — | — | — | — | — | |
| OTTERArchitecture=Cross-Encoder, Backbone=RemBERT2026.01 | 35.4 | — | — | — | — | — | — | — | — | |
| OTTERArchitecture=Cross-Encoder, Backbone=mmBERT2026.01 | 32.5 | — | — | — | — | — | — | — | — | |
| GLiNER-multi-v2.12026.01 | 29.1 | — | — | — | — | — | — | — | — | |
| OTTERArchitecture=Bi-Encoder, Backbone=mmBERT, Training Steps=100k2026.01 | 28.1 | — | — | — | — | — | — | — | — | |
| OTTERArchitecture=Bi-Encoder, Backbone=mmBERT2026.01 | 23.7 | — | — | — | — | — | — | — | — | |
| GPT-52026.01 | 20.4 | — | — | — | — | — | — | — | — | |
| GLiNER-x-base2026.01 | 18.7 | — | — | — | — | — | — | — | — | |
| OTTERArchitecture=Bi-Encoder, Backbone=RemBERT2026.01 | 16.6 | — | — | — | — | — | — | — | — | |
| WikiNeural2026.01 | 0.1 | — | — | — | — | — | — | — | — | |
| BERT (XLM-RoBERTa)Learning Scenario=Base-Supervised, Backbone=XLM-RoBERTa2024.09 | — | 41.9 | 33.5 | 29.1 | 23.4 | 32.9 | 29.2 | 27.2 | 28.6 | |
| CascadeNER-1.5BLearning Scenario=Base-Supervised, Backbone=Qwen2.5-1.5B2024.09 | — | 67.6 | 59.9 | 57.9 | 55.7 | 53.5 | 64 | 58.5 | 55.1 | |
| CascadeNER-7BLearning Scenario=Base-Supervised, Backbone=Qwen2.5-7B2024.09 | — | 73.8 | 65.5 | 60.3 | 59.6 | 61.4 | 69.8 | 65.3 | 62.7 | |
| CascadeNER-GPTLearning Scenario=Base-Supervised, Backbone=GPT-4o2024.09 | — | 77.1 | 71.7 | 69.9 | 70.3 | 70.8 | 74.3 | 72.4 | 70.9 | |
| GPT-NER-1.5BLearning Scenario=Base-Supervised, Backbone=Qwen2.5-1.5B2024.09 | — | 50.2 | 43.5 | 40.4 | 39.8 | 39.3 | 44.1 | 38.9 | 38.7 | |
| GPT-NER-7BLearning Scenario=Base-Supervised, Backbone=Qwen2.5-7B2024.09 | — | 55.1 | 48.2 | 47.2 | 44 | 48.1 | 50.9 | 44.8 | 44.5 | |
| GPT-NER-GPTLearning Scenario=Base-Supervised, Backbone=GPT-4o2024.09 | — | 62.4 | 58.3 | 57.9 | 56.8 | 56.9 | 60.4 | 57.3 | 55.9 | |
| PromptNER-1.5BLearning Scenario=Base-Supervised, Backbone=Qwen2.5-1.5B2024.09 | — | 21.6 | 18.6 | 17.1 | 14.9 | 15.8 | 20.7 | 16.4 | 15.9 | |
| PromptNER-7BLearning Scenario=Base-Supervised, Backbone=Qwen2.5-7B2024.09 | — | 41.1 | 32.9 | 31 | 30.7 | 30.3 | 47.4 | 35.6 | 29.6 | |
| PromptNER-GPTLearning Scenario=Base-Supervised, Backbone=GPT-4o2024.09 | — | 49.7 | 47.7 | 48.2 | 45.9 | 46.6 | 48.6 | 45.7 | 45.4 |