Relation Extraction on T-REx
87.7AccuracyRe2G
Evaluation Results
| Method | Links | |
|---|---|---|
| Re2GKnowledge Usage=With Knowledge, Evaluation Protocol=Fine-tuned2024.04 | 87.7 | |
| ODAKnowledge Usage=With Knowledge, Evaluation Protocol=Zero-Shot, Backbone LLM=GPT-42024.04 | 86 | |
| TOGKnowledge Usage=With Knowledge, Evaluation Protocol=Zero-Shot, Backbone LLM=GPT-42024.04 | 77.1 | |
| TOG-RKnowledge Usage=With Knowledge, Evaluation Protocol=Zero-Shot, Backbone LLM=GPT-42024.04 | 75.5 | |
| Direct answering (GPT4)Knowledge Usage=Without Knowledge, Evaluation Protocol=Zero-Shot, Backbone LLM=GPT-42024.04 | 57.72 | |
| Self-Consistency (GPT3.5)Knowledge Usage=Without Knowledge, Evaluation Protocol=Zero-Shot, Backbone LLM=GPT-3.52024.04 | 41.8 | |
| Direct answering (GPT3.5)Knowledge Usage=Without Knowledge, Evaluation Protocol=Zero-Shot, Backbone LLM=GPT-3.52024.04 | 37.78 | |
| COT (GPT3.5)Knowledge Usage=Without Knowledge, Evaluation Protocol=Zero-Shot, Backbone LLM=GPT-3.52024.04 | 32 |