Update README.md

README.md

@@ -129,6 +129,31 @@ The models have been fine-tuned on the following datasets.
 |        |[Synthetic-JP-EN-Coding-Dataset-567k](https://huggingface.co/datasets/Aratako/Synthetic-JP-EN-Coding-Dataset-567k)| A synthetic instruction dataset. We used sampled one.|
 |English |[FLAN](https://huggingface.co/datasets/Open-Orca/FLAN) | We used sampled one. | 
 
+
+## Evaluation
+
+### llm-jp-eval (v1.3.1)
+
+| Model name | average | EL | FA | HE | MC | MR | MT | NLI | QA | RC |
+| :--- | ---: | ---: | ---: | ---: | ---: | ---: | ---: | ---: | ---: | ---: |
+| [llm-jp-3-1.8b](https://huggingface.co/llm-jp/llm-jp-3-1.8b) | 0.3767 | 0.3725 | 0.1948 | 0.2350 | 0.2500 | 0.0900 | 0.7730 | 0.3080 | 0.4629 | 0.7040 |
+| [llm-jp-3-1.8b-instruct](https://huggingface.co/llm-jp/llm-jp-3-1.8b-instruct) | 0.4967 | 0.3695 | 0.1892 | 0.3850 | 0.4200 | 0.4200 | 0.7989 | 0.3700 | 0.5016 | 0.7729 |
+| [llm-jp-3-3.7b](https://huggingface.co/llm-jp/llm-jp-3-3.7b) | 0.4231 | 0.3812 | 0.2440 | 0.2200 | 0.1900 | 0.3600 | 0.7947 | 0.3800 | 0.4688 | 0.7694 |
+| [llm-jp-3-3.7b-instruct](https://huggingface.co/llm-jp/llm-jp-3-3.7b-instruct) | 0.5258 | 0.4220 | 0.2418 | 0.3950 | 0.5900 | 0.5600 | 0.8088 | 0.4260 | 0.4765 | 0.8123 |
+| [llm-jp-3-13b](https://huggingface.co/llm-jp/llm-jp-3-13b) | 0.5802 | 0.5570 | 0.2593 | 0.4600 | 0.7000 | 0.6300 | 0.8292 | 0.3460 | 0.5937 | 0.8469 |
+| [llm-jp-3-13b-instruct](https://huggingface.co/llm-jp/llm-jp-3-13b-instruct) | 0.6239 | 0.4848 | 0.2622 | 0.5300 | 0.9300 | 0.7000 | 0.8262 | 0.4460 | 0.5703 | 0.8659 |
+
+
+### Japanese MT Bench
+
+| Model name | average | coding | extraction | humanities | math | reasoning | roleplay | stem | writing |
+| :--- | ---: | ---: | ---: | ---: | ---: | ---: | ---: | ---: | ---: |
+| [llm-jp-3-1.8b-instruct](https://huggingface.co/llm-jp/llm-jp-3-1.8b-instruct) | 4.64 | 2.80 | 3.55 | 7.05 | 2.45 | 2.80 | 6.90 | 5.40 | 6.20 |
+| [llm-jp-3-3.7b-instruct](https://huggingface.co/llm-jp/llm-jp-3-3.7b-instruct) | 5.28 | 2.75 | 5.65 | 7.05 | 2.25 | 3.65 | 8.25 | 5.85 | 6.75 |
+| [llm-jp-3-13b-instruct](https://huggingface.co/llm-jp/llm-jp-3-13b-instruct) | 6.66 | 4.30 | 6.90 | 9.00 | 3.60 | 5.75 | 8.55 | 7.50 | 7.70 |
+
+
+
 ## Risks and Limitations
 
 The models released here are in the early stages of our research and development and have not been tuned to ensure outputs align with human intent and safety considerations.