Files changed (1) hide show
  1. README.md +7 -3
README.md CHANGED
@@ -54,9 +54,13 @@ Please take it with a pinch of salt as we continue to study Modello Italia.
54
  * The model architecture is **based on GPT-NeoX**.
55
 
56
  ## Results
57
- Modello Italia 9B has not been evaluated on standard benchmarks yet.
58
- We will update this model card with the results soon.
59
- * **Want to contribute to the evaluation?** Submit a pull request!
 
 
 
 
60
 
61
  ## How to use Modello Italia with Hugging Face transformers
62
 
 
54
  * The model architecture is **based on GPT-NeoX**.
55
 
56
  ## Results
57
+ For a detailed comparison of model performance, check out the [Leaderboard for Italian Language Models](https://huggingface.co/spaces/FinancialSupport/open_ita_llm_leaderboard).
58
+
59
+ Here's a breakdown of the performance metrics:
60
+
61
+ | Metric | hellaswag_it acc_norm | arc_it acc_norm | m_mmlu_it 5-shot acc | Average |
62
+ |:----------------------------|:----------------------|:----------------|:---------------------|:--------|
63
+ | **Accuracy Normalized** | 0.5679 | 0.3849 | 0.3522 | 0.4350 |
64
 
65
  ## How to use Modello Italia with Hugging Face transformers
66