Commit
·
73cdbde
1
Parent(s):
151e2f3
Adding Evaluation Results (#2)
Browse files- Adding Evaluation Results (5ef233527fdf12bfcad8326a70276b3836247bd7)
Co-authored-by: Open LLM Leaderboard PR Bot <leaderboard-pr-bot@users.noreply.huggingface.co>
README.md
CHANGED
|
@@ -22,4 +22,17 @@ A little bit dumb, but good for creative scenarios.
|
|
| 22 |
Note the whitespace - the prefixes for messages are `" ***System:"`, `" ***Query:"`, and `" ***Response:"`. This is important as `"***"` and `" ***"` are two entirely different tokens.
|
| 23 |
|
| 24 |
|
| 25 |
-
[<img src="https://raw.githubusercontent.com/OpenAccess-AI-Collective/axolotl/main/image/axolotl-badge-web.png" alt="Built with Axolotl" width="200" height="32"/>](https://github.com/OpenAccess-AI-Collective/axolotl)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 22 |
Note the whitespace - the prefixes for messages are `" ***System:"`, `" ***Query:"`, and `" ***Response:"`. This is important as `"***"` and `" ***"` are two entirely different tokens.
|
| 23 |
|
| 24 |
|
| 25 |
+
[<img src="https://raw.githubusercontent.com/OpenAccess-AI-Collective/axolotl/main/image/axolotl-badge-web.png" alt="Built with Axolotl" width="200" height="32"/>](https://github.com/OpenAccess-AI-Collective/axolotl)
|
| 26 |
+
# [Open LLM Leaderboard Evaluation Results](https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard)
|
| 27 |
+
Detailed results can be found [here](https://huggingface.co/datasets/open-llm-leaderboard/details_chargoddard__ypotryll-22b-epoch2-qlora)
|
| 28 |
+
|
| 29 |
+
| Metric | Value |
|
| 30 |
+
|-----------------------|---------------------------|
|
| 31 |
+
| Avg. | 51.68 |
|
| 32 |
+
| ARC (25-shot) | 59.22 |
|
| 33 |
+
| HellaSwag (10-shot) | 80.66 |
|
| 34 |
+
| MMLU (5-shot) | 54.52 |
|
| 35 |
+
| TruthfulQA (0-shot) | 40.42 |
|
| 36 |
+
| Winogrande (5-shot) | 76.32 |
|
| 37 |
+
| GSM8K (5-shot) | 5.38 |
|
| 38 |
+
| DROP (3-shot) | 45.24 |
|