Conversation Forecasting on Conversation Forecasting Datasets In-domain
19.6BSFortune Dial (IF x4)
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Fortune Dial (IF x4)Model Backbone=Llama-2-chat 7B, Forecasting Strategy=Implicit Forecasts, Data Scale=x42024.02 | 19.6 | 11.4 | |
| Fortune Dial (IF)Model Backbone=Llama-2-chat 7B, Forecasting Strategy=Implicit Forecasts, Data Scale=standard2024.02 | 21 | 5.3 | |
| Fortune Dial (IF)Model Backbone=Llama-2 1B, Forecasting Strategy=Implicit Forecasts, Data Scale=standard2024.02 | 22.2 | -0.5 | |
| Fortune Dial (IF)Model Backbone=Zephyr 7B, Forecasting Strategy=Implicit Forecasts, Data Scale=standard2024.02 | 22.5 | -1.6 | |
| Fortune Dial (DF)Model Backbone=Llama-2-chat 7B, Forecasting Strategy=Direct Forecasts, Data Scale=standard2024.02 | 22.8 | -3.2 | |
| Fortune Dial (DF x4)Model Backbone=Llama-2-chat 7B, Forecasting Strategy=Direct Forecasts, Data Scale=x42024.02 | 22.8 | -3.2 | |
| Fortune Dial (DF rl)Model Backbone=Llama-2-chat 7B, Forecasting Strategy=Direct Forecasts (RL), Data Scale=standard2024.02 | 22.9 | -3.7 | |
| Fortune Dial (DF rl)Model Backbone=Zephyr 7B, Forecasting Strategy=Direct Forecasts (RL), Data Scale=standard2024.02 | 23 | -4.2 | |
| Fortune Dial (DF rl)Model Backbone=Llama-2 1B, Forecasting Strategy=Direct Forecasts (RL), Data Scale=standard2024.02 | 23 | -4.2 |