Adding Evaluation Results
#1
by
leaderboard-pr-bot
- opened
README.md
CHANGED
|
@@ -55,3 +55,17 @@ Thank you to LAION and Stability.ai for support and compute resources to experim
|
|
| 55 |
- Conditional pretraining helps the user control the outputs of the model.
|
| 56 |
- However, these models (and all language models) can still generate undesirable content.
|
| 57 |
- So please enjoy and use with care!
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 55 |
- Conditional pretraining helps the user control the outputs of the model.
|
| 56 |
- However, these models (and all language models) can still generate undesirable content.
|
| 57 |
- So please enjoy and use with care!
|
| 58 |
+
|
| 59 |
+
# [Open LLM Leaderboard Evaluation Results](https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard)
|
| 60 |
+
Detailed results can be found [here](https://huggingface.co/datasets/open-llm-leaderboard/details_Rallio67__3B-redpajama-conditional-alpha)
|
| 61 |
+
|
| 62 |
+
| Metric | Value |
|
| 63 |
+
|-----------------------|---------------------------|
|
| 64 |
+
| Avg. | 32.37 |
|
| 65 |
+
| ARC (25-shot) | 36.26 |
|
| 66 |
+
| HellaSwag (10-shot) | 61.9 |
|
| 67 |
+
| MMLU (5-shot) | 25.42 |
|
| 68 |
+
| TruthfulQA (0-shot) | 36.31 |
|
| 69 |
+
| Winogrande (5-shot) | 60.77 |
|
| 70 |
+
| GSM8K (5-shot) | 0.61 |
|
| 71 |
+
| DROP (3-shot) | 5.28 |
|