Factual Precision Evaluation on Culture & Entertainment
0.887FACTSCOREMistralINST
Evaluation Results
| Method | Links | |
|---|---|---|
| MistralINSTScenario=INFO, CORE=w/o2024.07 | 0.887 | |
| MistralINSTScenario=REP, CORE=w/o2024.07 | 0.879 | |
| GPT-2Scenario=INFO, CORE=w/o2024.07 | 0.871 | |
| MistralINSTScenario=NORMAL, CORE=w/o2024.07 | 0.815 | |
| MistralINSTScenario=NORMAL, CORE=w/2024.07 | 0.797 | |
| GPT-2Scenario=REP, CORE=w/o2024.07 | 0.727 | |
| MistralINSTScenario=INFO, CORE=w/2024.07 | 0.404 | |
| MistralINSTScenario=REP, CORE=w/2024.07 | 0.264 | |
| GPT-2Scenario=INFO, CORE=w/2024.07 | 0.0341 | |
| GPT-2Scenario=REP, CORE=w/2024.07 | 0.01 |