Adding Evaluation Results
#2
by leaderboard-pr-bot - opened
README.md
CHANGED
|
@@ -114,4 +114,17 @@ unfiltered content from the internet, which is far from neutral. As the openAI t
|
|
| 114 |
> not recommend that they be deployed into systems that interact with humans unless the deployers first carry out a
|
| 115 |
> study of biases relevant to the intended use-case. We found no statistically significant difference in gender, race,
|
| 116 |
> and religious bias probes between 774M and 1.5B, implying all versions of GPT-2 should be approached with similar
|
| 117 |
-
> levels of caution around use cases that are sensitive to biases around human attributes.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 114 |
> not recommend that they be deployed into systems that interact with humans unless the deployers first carry out a
|
| 115 |
> study of biases relevant to the intended use-case. We found no statistically significant difference in gender, race,
|
| 116 |
> and religious bias probes between 774M and 1.5B, implying all versions of GPT-2 should be approached with similar
|
| 117 |
+
> levels of caution around use cases that are sensitive to biases around human attributes.
|
| 118 |
+
# [Open LLM Leaderboard Evaluation Results](https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard)
|
| 119 |
+
Detailed results can be found [here](https://huggingface.co/datasets/open-llm-leaderboard/details_crumb__gpt2023)
|
| 120 |
+
|
| 121 |
+
| Metric | Value |
|
| 122 |
+
|-----------------------|---------------------------|
|
| 123 |
+
| Avg. | 24.85 |
|
| 124 |
+
| ARC (25-shot) | 21.93 |
|
| 125 |
+
| HellaSwag (10-shot) | 31.11 |
|
| 126 |
+
| MMLU (5-shot) | 25.05 |
|
| 127 |
+
| TruthfulQA (0-shot) | 40.71 |
|
| 128 |
+
| Winogrande (5-shot) | 50.12 |
|
| 129 |
+
| GSM8K (5-shot) | 0.3 |
|
| 130 |
+
| DROP (3-shot) | 4.73 |
|