Language Modeling on StackExchange (val)
4.43PerplexityStackLLaMA (Initial)
Evaluation results
| Method | Links | |
|---|---|---|
| StackLLaMA (Initial)Mode=Zero-shot2023.12 | 4.43 | |
| Elastic ResetTraining Protocol=RLHF, Epochs=6002023.12 | 4.57 | |
| PPOTraining Protocol=RLHF, Epochs=6002023.12 | 4.62 |