Adding Evaluation Results
#16
by
leaderboard-pr-bot
- opened
README.md
CHANGED
@@ -251,3 +251,17 @@ And thank you again to a16z for their generous grant.
|
|
251 |
# Original model card: Tim Dettmers' Guanaco 33B
|
252 |
|
253 |
No original model card was available.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
251 |
# Original model card: Tim Dettmers' Guanaco 33B
|
252 |
|
253 |
No original model card was available.
|
254 |
+
|
255 |
+
# [Open LLM Leaderboard Evaluation Results](https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard)
|
256 |
+
Detailed results can be found [here](https://huggingface.co/datasets/open-llm-leaderboard/details_TheBloke__guanaco-33B-GPTQ)
|
257 |
+
|
258 |
+
| Metric | Value |
|
259 |
+
|-----------------------|---------------------------|
|
260 |
+
| Avg. | 33.87 |
|
261 |
+
| ARC (25-shot) | 28.16 |
|
262 |
+
| HellaSwag (10-shot) | 26.34 |
|
263 |
+
| MMLU (5-shot) | 24.94 |
|
264 |
+
| TruthfulQA (0-shot) | 48.98 |
|
265 |
+
| Winogrande (5-shot) | 78.85 |
|
266 |
+
| GSM8K (5-shot) | 23.81 |
|
267 |
+
| DROP (3-shot) | 5.98 |
|