Commit
•
fca423b
1
Parent(s):
b57b9f5
Adding Evaluation Results (#7)
Browse files- Adding Evaluation Results (e7128026c902277c6664dd527ac61d251b75dc3c)
Co-authored-by: Open LLM Leaderboard PR Bot <[email protected]>
README.md
CHANGED
@@ -74,4 +74,17 @@ The model may not perform effectively outside the scope of the medical domain.
|
|
74 |
The training data primarily targets the knowledge level of medical students,
|
75 |
which may result in limitations when addressing the needs of board-certified physicians.
|
76 |
The model has not been tested in real-world applications, so its efficacy and accuracy are currently unknown.
|
77 |
-
It should never be used as a substitute for a doctor's opinion and must be treated as a research tool only.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
74 |
The training data primarily targets the knowledge level of medical students,
|
75 |
which may result in limitations when addressing the needs of board-certified physicians.
|
76 |
The model has not been tested in real-world applications, so its efficacy and accuracy are currently unknown.
|
77 |
+
It should never be used as a substitute for a doctor's opinion and must be treated as a research tool only.
|
78 |
+
# [Open LLM Leaderboard Evaluation Results](https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard)
|
79 |
+
Detailed results can be found [here](https://huggingface.co/datasets/open-llm-leaderboard/details_medalpaca__medalpaca-7b)
|
80 |
+
|
81 |
+
| Metric | Value |
|
82 |
+
|-----------------------|---------------------------|
|
83 |
+
| Avg. | 44.98 |
|
84 |
+
| ARC (25-shot) | 54.1 |
|
85 |
+
| HellaSwag (10-shot) | 80.42 |
|
86 |
+
| MMLU (5-shot) | 41.47 |
|
87 |
+
| TruthfulQA (0-shot) | 40.46 |
|
88 |
+
| Winogrande (5-shot) | 71.19 |
|
89 |
+
| GSM8K (5-shot) | 3.03 |
|
90 |
+
| DROP (3-shot) | 24.21 |
|