Conversational Question Answering on ConvFinQA (dev)
43.41Exact Match AccBloombergGPT
Evaluation Results
| Method | Links | |
|---|---|---|
| BloombergGPTInput context=Entire gold conversation2023.03 | 43.41 | |
| BLOOM176BInput context=Entire gold conversation2023.03 | 36.31 | |
| GPT-NeoXInput context=Entire gold conversation2023.03 | 30.06 | |
| OPT66BInput context=Entire gold conversation2023.03 | 27.88 |