Named Entity Recognition on NER
75.5Hard F1 ScoreTAG
Evaluation results
| Method | Links | ||
|---|---|---|---|
| TAGBackbone=5-m2026.01 | 75.5 | — | |
| MATCHBackbone=70B2026.01 | 72.7 | — | |
| MATCH-SBackbone=70B2026.01 | 72.4 | — | |
| LOGITMATCHBackbone=70B2026.01 | 72.4 | — | |
| LOGITMATCH-SBackbone=70B2026.01 | 72.3 | — | |
| LOGITMATCH-SBackbone=8B2026.01 | 71.7 | — | |
| MATCH-OCC-SBackbone=8B2026.01 | 71.5 | — | |
| LOGITMATCH-OCCBackbone=8B2026.01 | 71.5 | — | |
| LOGITMATCHBackbone=8B2026.01 | 71.4 | — | |
| MATCH-SBackbone=5-m2026.01 | 71.3 | — | |
| LOGITMATCH-OCC-SBackbone=8B2026.01 | 71.3 | — | |
| MATCH-OCCBackbone=8B2026.01 | 71.1 | — | |
| MATCH-SBackbone=8B2026.01 | 71 | — | |
| MATCHBackbone=8B2026.01 | 70.9 | — | |
| TAGBackbone=70B2026.01 | 70.4 | — | |
| INDEX-ENRICHEDBackbone=5-m2026.01 | 69.1 | — | |
| LOGITMATCH-OCCBackbone=70B2026.01 | 68.3 | — | |
| LOGITMATCH-OCC-SBackbone=70B2026.01 | 68.3 | — | |
| MATCH-OCCBackbone=70B2026.01 | 68.2 | — | |
| MATCH-OCC-SBackbone=70B2026.01 | 67.3 | — | |
| MATCH-OCCBackbone=5-m2026.01 | 66.7 | — | |
| MATCH-OCC-SBackbone=5-m2026.01 | 65 | — | |
| TAGBackbone=8B2026.01 | 63.4 | — | |
| LOGITMATCH-OCCBackbone=24B2026.01 | 63.4 | — | |
| TAGBackbone=24B2026.01 | 63.3 | — | |
| MATCH-OCCBackbone=24B2026.01 | 63 | — | |
| INDEX-ENRICHEDBackbone=24B2026.01 | 61.6 | — | |
| INDEX-ENRICHEDBackbone=70B2026.01 | 59.3 | — | |
| MATCHBackbone=5-m2026.01 | 59 | — | |
| LOGITMATCHBackbone=24B2026.01 | 53.7 | — | |
| MATCH-OCC-SBackbone=24B2026.01 | 53.6 | — | |
| MATCHBackbone=24B2026.01 | 53.5 | — | |
| LOGITMATCH-OCC-SBackbone=24B2026.01 | 53 | — | |
| MATCH-SBackbone=24B2026.01 | 49.1 | — | |
| LOGITMATCH-SBackbone=24B2026.01 | 48.6 | — | |
| INDEX-ENRICHEDBackbone=8B2026.01 | 39.6 | — | |
| INDEXBackbone=5-m2026.01 | 33.8 | — | |
| INDEXBackbone=70B2026.01 | 23.3 | — | |
| INDEXBackbone=8B2026.01 | 17.8 | — | |
| INDEXBackbone=24B2026.01 | 16.7 | — | |
| baselineClassifier=CRF2025.04 | — | 71.5 | |
| BERTParams=110M2023.10 | — | 75.09 | |
| BERT2023.10 | — | 75.09 | |
| BERT-MULTIClassifier=MLP2025.04 | — | 87.04 | |
| BERT-MULTIClassifier=CRF2025.04 | — | 87.28 | |
| BLOOM176BShots=202023.03 | — | 55.56 | |
| BloombergGPTShots=202023.03 | — | 60.82 | |
| BloombergGPTParams=50B2023.10 | — | 60.82 | |
| FiLMParams=110M, Training tokens=2.4B2023.10 | — | 79.79 | |
| FiLM# Tokens (Financial Corpus)=2.4B2023.10 | — | 79.78 | |
| FiLM (5.5B)# Tokens (Financial Corpus)=5.5B2023.10 | — | 78.82 | |
| FinBERT-A# Tokens (Financial Corpus)=237M2023.10 | — | 77.93 | |
| FinBERT-Y# Tokens (Financial Corpus)=4.9B2023.10 | — | 70.4 | |
| FLANG-BERT# Tokens (Financial Corpus)=NA2023.10 | — | 75.58 | |
| FLANG-ROBERTa# Tokens (Financial Corpus)=NA2023.10 | — | 71.36 | |
| GPT-NeoModel Size=125M, k=102022.11 | — | 15.52 | |
| GPT-NeoModel Size=1.3B, k=102022.11 | — | 39.07 | |
| GPT-NeoModel Size=2.7B, k=102022.11 | — | 40.16 | |
| GPT-NeoModel Size=6B, k=102022.11 | — | 53.03 | |
| GPT-NeoModel Size=20B, k=102022.11 | — | 58.05 | |
| GPT-NeoModel Size=20B, k=02022.11 | — | 3.79 | |
| GPT-NeoModel Size=20B, k=12022.11 | — | 10.26 | |
| GPT-NeoModel Size=20B, k=32022.11 | — | 33.63 | |
| GPT-NeoModel Size=20B, k=52022.11 | — | 40.6 | |
| GPT-NeoModel Size=20B, k=72022.11 | — | 52.12 | |
| GPT-NeoModel Size=20B, k=92022.11 | — | 58.08 | |
| GPT-NeoXShots=202023.03 | — | 60.98 | |
| KLUE-ROBERTAClassifier=MLP2025.04 | — | 88.77 | |
| KLUE-ROBERTAClassifier=CRF2025.04 | — | 88.84 | |
| LSTMClassifier=CRF, Word Embedding=fastText2025.04 | — | 84.76 | |
| Majority Label2022.11 | — | 0 | |
| OPT66BShots=202023.03 | — | 57.49 | |
| Per-Span Majority2022.11 | — | 47.52 | |
| RoBERTaParams=110M2023.10 | — | 78.81 | |
| ROBERTa2023.10 | — | 78.81 | |
| SEC-BERT# Tokens (Financial Corpus)=3.1B2023.10 | — | 78.74 | |
| XLM-ROBERTAClassifier=MLP2025.04 | — | 87.93 | |
| XLM-ROBERTAClassifier=CRF2025.04 | — | 88.16 |