Update README.md

README.md

@@ -13,8 +13,28 @@ tags:
 Converted version of [Mixtral Instruct](https://huggingface.co/mistralai/Mixtral-8x7B-Instruct-v0.1) to 4-bit using bitsandbytes. For more information about the model, refer to the model's page.
 
 ### Impact on performance
-Impact of quantization on a set of models. The impact of quantization is negligible.
-![constellation](https://i.postimg.cc/T20mBPxD/constellation.png)
+Impact of quantization on a set of models.
+
+Evaluation of the model was conducted using the PoLL (Pool of LLM) technique, assessing performance on **100 French questions** with scores aggregated from six evaluations
+(two per evaluator). The evaluators included GPT-4o, Gemini-1.5-pro, and Claude3.5-sonnet.
+
+**Performance Scores (on a scale of 5):**
+| Model                                        | Score   | # params |
+|---------------------------------------------:|:-------:|:--------:|
+| gpt-4o                                       | 4.13    | N/A      |
+| mistralai/Mixtral-8x7B-Instruct-v0.1         | 3.71    | 46.7b    |
+| cmarkea/Mixtral-8x7B-Instruct-v0.1-4bit      | 3.68    | 46.7b    |
+| gpt-3.5-turbo                                | 3.66    | 175b     |
+| TheBloke/Mixtral-8x7B-Instruct-v0.1-GPTQ     | 3.56    | 46.7b    |
+| mistralai/Mistral-7B-Instruct-v0.2           | 1.98    | 7.25b    |
+| cmarkea/bloomz-7b1-mt-sft-chat               | 1.69    | 7.1b     |
+| cmarkea/bloomz-3b-dpo-chat                   | 1.68    | 3b       |
+| cmarkea/bloomz-3b-sft-chat                   | 1.51    | 3b       |
+| croissantllm/CroissantLLMChat-v0.1           | 1.19    | 1.3b     |
+| cmarkea/bloomz-560m-sft-chat                 | 1.04    | 0.56b    |
+| OpenLLM-France/Claire-Mistral-7B-0.1         | 0.38    | 7.25b    |
+
+The impact of quantization is negligible.
 
 ### Prompt Pattern
 Here is a reminder of the command pattern to interact with the model if the add_special_tokens option is disabled (otherwise, do not include the BOS symbol and space at the beginning of the sequence):