Update README.md
Browse files
README.md
CHANGED
|
@@ -13,6 +13,12 @@ metrics:
|
|
| 13 |
- spearmanr
|
| 14 |
- accuracy
|
| 15 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 16 |
|
| 17 |
# TL;DR
|
| 18 |
Prometheus is an alternative of GPT-4 evaluation when doing fine-grained evaluation of an underlying LLM & a Reward model for Reinforcement Learning from Human Feedback (RLHF).
|
|
|
|
| 13 |
- spearmanr
|
| 14 |
- accuracy
|
| 15 |
---
|
| 16 |
+
## Links for Reference
|
| 17 |
+
|
| 18 |
+
- **Homepage:https://github.com/kaistAI/Prometheus**
|
| 19 |
+
- **Repository:https://github.com/kaistAI/Prometheus**
|
| 20 |
+
- **Paper:https://arxiv.org/abs/2310.08491**
|
| 21 |
+
- **Point of Contact:seungone@kaist.ac.kr**
|
| 22 |
|
| 23 |
# TL;DR
|
| 24 |
Prometheus is an alternative of GPT-4 evaluation when doing fine-grained evaluation of an underlying LLM & a Reward model for Reinforcement Learning from Human Feedback (RLHF).
|