Adding Evaluation Results
#2
by
leaderboard-pr-bot
- opened
README.md
CHANGED
|
@@ -12,4 +12,17 @@ widget:
|
|
| 12 |
|
| 13 |
This is a follow up finetune of Pythia-70M finetuned on the same dataset as OPT-19M-ChatSalad. It is much more coherent.
|
| 14 |
|
| 15 |
-
All feedback and comments can be directed to Concedo on the KoboldAI discord.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 12 |
|
| 13 |
This is a follow up finetune of Pythia-70M finetuned on the same dataset as OPT-19M-ChatSalad. It is much more coherent.
|
| 14 |
|
| 15 |
+
All feedback and comments can be directed to Concedo on the KoboldAI discord.
|
| 16 |
+
# [Open LLM Leaderboard Evaluation Results](https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard)
|
| 17 |
+
Detailed results can be found [here](https://huggingface.co/datasets/open-llm-leaderboard/details_concedo__Pythia-70M-ChatSalad)
|
| 18 |
+
|
| 19 |
+
| Metric | Value |
|
| 20 |
+
|-----------------------|---------------------------|
|
| 21 |
+
| Avg. | 25.15 |
|
| 22 |
+
| ARC (25-shot) | 20.99 |
|
| 23 |
+
| HellaSwag (10-shot) | 27.28 |
|
| 24 |
+
| MMLU (5-shot) | 24.78 |
|
| 25 |
+
| TruthfulQA (0-shot) | 49.74 |
|
| 26 |
+
| Winogrande (5-shot) | 52.41 |
|
| 27 |
+
| GSM8K (5-shot) | 0.0 |
|
| 28 |
+
| DROP (3-shot) | 0.84 |
|