Adding Evaluation Results
#3
by
leaderboard-pr-bot
- opened
README.md
CHANGED
|
@@ -306,3 +306,17 @@ Training took about 1 hour using QLoRa on 1xA100, so this model can be recreated
|
|
| 306 |
- A bit more data with more diversity.
|
| 307 |
- Refactor dataset generation script.
|
| 308 |
- Test some model merges using this model.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 306 |
- A bit more data with more diversity.
|
| 307 |
- Refactor dataset generation script.
|
| 308 |
- Test some model merges using this model.
|
| 309 |
+
|
| 310 |
+
# [Open LLM Leaderboard Evaluation Results](https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard)
|
| 311 |
+
Detailed results can be found [here](https://huggingface.co/datasets/open-llm-leaderboard/details_TheBloke__EverythingLM-13B-16K-GPTQ)
|
| 312 |
+
|
| 313 |
+
| Metric | Value |
|
| 314 |
+
|-----------------------|---------------------------|
|
| 315 |
+
| Avg. | 30.29 |
|
| 316 |
+
| ARC (25-shot) | 29.27 |
|
| 317 |
+
| HellaSwag (10-shot) | 26.24 |
|
| 318 |
+
| MMLU (5-shot) | 25.4 |
|
| 319 |
+
| TruthfulQA (0-shot) | 48.58 |
|
| 320 |
+
| Winogrande (5-shot) | 71.35 |
|
| 321 |
+
| GSM8K (5-shot) | 5.38 |
|
| 322 |
+
| DROP (3-shot) | 5.83 |
|