Update README.md

README.md

@@ -20,9 +20,7 @@ datasets:
 # Model description
 Heidrun-Mistral-7B-chat is a chat-model based on [Heidrun-Mistral-7B-base](https://huggingface.co/Mabeck/Heidrun-Mistral-7B-base), finetuned on [danish-OpenHermes](https://huggingface.co/datasets/Mabeck/danish-OpenHermes) and [skoleGPT](https://huggingface.co/datasets/kobprof/skolegpt-instruct) for a instruction/chat format.
 
-It achieved 35.66%+-0.85%/51.68+-0.63% on MMLU-Da in [ScandEval](https://github.com/ScandEval/ScandEval) which is the best non-merged opensource model score to date (13/02/24).
-
-Further evaluations will be tested.
+It is a new SOTA Danish open-source LLM and shows very strong performance in logic and reasoning tasks.
 
 # Benchmarks
 
@@ -32,6 +30,8 @@ The following benchmarks have been tested using [ScandEval](https://github.com/S
 - **DANSK**: 50.77%+-2.29%/34.05%+-1.78%, ranks 3rd=
 - **Hellaswag-da**: 29.18%+-0.99%/46.64%+-0.76%, ranks 4th
 
+Further evaluations will be tested.
+
 # Datasets
 This model is trained on Danish instruction datasets [danish-OpenHermes](Mabeck/danish-OpenHermes) and [skoleGPT](https://huggingface.co/datasets/kobprof/skolegpt-instruct), which have not been safeguarded or alligned.