Named Entity Recognition on Cross-domain NER datasets out-of-domain
62.9AI NER ScoreUniNER-7B
Evaluation results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| UniNER-7Bin-domain supervised=true, variant=inst-tuned + sup.2023.08 | 62.9 | 61.2 | 35.2 | 64.9 | 70.6 | 66.9 | 70.8 | 61.8 | |
| GoldFTParams=8B, Backbone=Llama 3.1, Evaluation Protocol=Fine-tuned2025.05 | 62.6 | 63.6 | 54 | 65.4 | 67.4 | 66.3 | 60.1 | 62.8 | |
| GUIDEX_FT + GoldFTParams=8B, Backbone=Llama 3.1, Evaluation Protocol=Fine-tuned2025.05 | 62.4 | 65.3 | 55.5 | 63.8 | 67.9 | 69.6 | 64.6 | 64.2 | |
| UniNER-7Bin-domain supervised=true, variant=sup. only2023.08 | 62.3 | 54.2 | 16 | 67.4 | 69 | 64.5 | 66.9 | 57.2 | |
| KnowCoderParams=7B, Backbone=Llama 2, Evaluation Protocol=Zero-Shot2025.05 | 60.3 | 50 | 48.2 | 61.1 | 70 | 72.2 | 59.1 | 60.1 | |
| GOLLIEParams=7B, Backbone=CodeLlama, Evaluation Protocol=Zero-Shot2025.05 | 59.1 | 63 | 43.4 | 62.7 | 67.8 | 57.2 | 55.5 | 58 | |
| GLINER-LParams=0.3B, Backbone=DeBERTa-V3, Evaluation Protocol=Zero-Shot2025.05 | 57.2 | 57.2 | 42.9 | 64.4 | 69.6 | 72.6 | 62.6 | 60.9 | |
| UniNER-13Bzero-shot=true2023.08 | 54.2 | 48.7 | 36.2 | 60.9 | 64.5 | 61.4 | 63.5 | 55.6 | |
| UniNER-13BParams=13B, Backbone=Llama, Evaluation Protocol=Zero-Shot2025.05 | 54.2 | 48.7 | 36.2 | 60.9 | 64.5 | 61.4 | 63.5 | 55.6 | |
| UniNER-7BParams=7B, Backbone=Llama, Evaluation Protocol=Zero-Shot2025.05 | 53.6 | 42.4 | 31.7 | 59.3 | 67 | 60.9 | 61.1 | 53.7 | |
| UniNER-7Bzero-shot=true2023.08 | 53.5 | 42.4 | 31.7 | 59.4 | 65 | 60.8 | 61.1 | 53.4 | |
| ChatGPTzero-shot=true2023.08 | 52.4 | 5.3 | 32.8 | 39.8 | 66.6 | 68.5 | 67 | 47.5 | |
| ChatGPTEvaluation Protocol=Zero-Shot2025.05 | 52.4 | 5.3 | 32.8 | 39.8 | 66.6 | 68.5 | 67 | 47.5 | |
| InstructUIEParams=11B, Backbone=FlanT5, Evaluation Protocol=Zero-Shot2025.05 | 49 | 63 | 21 | 47.2 | 53.2 | 48.1 | 49.2 | 47.2 | |
| InstructUIE-11Bin-domain supervised=true2023.08 | 48.4 | — | — | 48.8 | 54.4 | 49.9 | 49.4 | — | |
| KnowCoder-7B (SU. only)Setting=few-shot, Shot count=5-shot, Training Phase=Schema Understanding (SU) only2024.03 | 41.8 | 37.2 | 36.4 | 42.6 | 53.8 | 60.6 | 51.6 | 46.3 | |
| LLAMA2-13BSetting=few-shot, Shot count=5-shot2024.03 | 37.5 | 32.6 | 25.2 | 36.5 | 37 | 60.3 | 51.7 | 40.1 | |
| LLaMA2-7BSetting=few-shot, Shot count=5-shot2024.03 | 30.8 | 31 | 19.6 | 24.1 | 28 | 38.7 | 44.1 | 30.9 | |
| USMParams=0.3B, Evaluation Protocol=Zero-Shot2025.05 | 28.2 | 37.7 | 17.7 | 56 | 44.9 | 36.1 | 44 | 37.8 | |
| Vicuna-13Bzero-shot=true2023.08 | 22.7 | 0.9 | 0.4 | 22.7 | 26.6 | 27.2 | 22 | 17.5 | |
| Vicuna-13BParams=13B, Backbone=Llama, Evaluation Protocol=Zero-Shot2025.05 | 22.7 | 0.9 | 0.4 | 22.7 | 26.6 | 27 | 22 | 17.5 | |
| Vicuna-7Bzero-shot=true2023.08 | 12.8 | 6 | 5.3 | 16.1 | 17 | 20.5 | 13 | 13 | |
| Vicuna-7BParams=7B, Backbone=Llama, Evaluation Protocol=Zero-Shot2025.05 | 12.8 | 6 | 5.3 | 16.1 | 17 | 20.5 | 13 | 13 |