Update README.md

README.md

@@ -28,6 +28,7 @@ SambaLingo-Russian-Base is a pretrained Bi-lingual Russian and English model tha
 - **Language(s):** Russian, English
 - **Finetuned from model:** [Llama 2](https://huggingface.co/meta-llama/Llama-2-7b-hf)
 - **Try the chat version of this model**: [SambaLingo-chat-space](https://huggingface.co/spaces/sambanovasystems/SambaLingo-chat-space).
+- **Paper:** [SambaLingo: Teaching Large Language Models New Languages](https://arxiv.org/abs/2404.05829)
 - **Blog Post**: [sambalingo-open-source-language-experts](https://sambanova.ai/blog/sambalingo-open-source-language-experts)
 
 ## Getting Started
@@ -53,19 +54,7 @@ All pre-training is done on the [Cultura-X](https://huggingface.co/datasets/uonl
 We extended the vocabulary of the base llama model from 32,000 tokens to 57,000 tokens by adding up to 25,000 non-overlapping tokens from the new language.
 
 ## Evaluation Results 
-|        | SambaLingo-Russian-Base | saiga_mistral_7b_merged | rugpt3large_based_on_gpt2 | bloom-7b1 | xglm-7.5B | mGPT-13B |
-|------------------------------------------|-----------------------------------|--------------------------------------|----------------------|--------------------|---------------------|--------|
-| Holdout Perplexity (Lower is better)     | **1.444**                             | 1.556                                | 1.611                | 1.797              | 1.504               | 1.806  |
-| FLORES en->ru (8 shot, CHRF)             | **0.472**                             | 0.425                                | 0.319                | 0.204              | 0.263               | 0.211  |
-| FLORES ru->en (8 shot, CHRF)             | **0.587**                             | 0.527                                | 0.317                | 0.258              | 0.429               | 0.251  |
-| FLORES en->ru  (8 shot, BLEU)            | **0.194**                             | 0.145                                | 0.074                | 0.012              | 0.045               | 0.021  |
-| FLORES ru->en  (8 shot, BLEU)            | **0.301**                             | 0.249                                | 0.062                | 0.032              | 0.152               | 0.039  |
-| Belebele (3 shot)                        | **39.00%**                            | 34.44%                               | 24.33%               | 29.00%             | 21.89%              | 23.67% |
-| SIB-200 (3 shot)                         | 69.12%                            | **78.92%**                               | 32.84%               | 46.08%             | 63.73%              | 42.65% |
-| XNLI (0 shot)                            | 35.29%                            | **49.78%**                               | 45.61%               | 42.61%             | 46.39%              | 45.39% |
-| XStoryCloze (0 shot)                     | **71.67%**                            | 68.96%                               | 60.75%               | 52.68%             | 63.40%              | 59.43% |
-| XWinograd (0 shot)                       | **69.21%**                            | 66.67%                               | 60.63%               | 57.14%             | 63.17%              | 60.00% |
-
+For evaluation results see our paper: [SambaLingo: Teaching Large Language Models New Languages](https://arxiv.org/abs/2404.05829)
 
 ## Uses
 <!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
@@ -107,12 +96,12 @@ We would like to give a special thanks to the following groups:
 
 ## Cite SambaLingo
 ```
-@software{sambalingo,
-  title = {{SambaLingo: Open Source Language Experts}},
-  author = {SambaNova Systems},
-  url = {https://huggingface.co/sambanovasystems/SambaLingo-Russian-Base}
-  month = {2},
-  year = {2024},
-  version = {1.0},
+@misc{csaki2024sambalingo,
+      title={SambaLingo: Teaching Large Language Models New Languages}, 
+      author={Zoltan Csaki and Bo Li and Jonathan Li and Qiantong Xu and Pian Pawakapan and Leon Zhang and Yun Du and Hengyu Zhao and Changran Hu and Urmish Thakker},
+      year={2024},
+      eprint={2404.05829},
+      archivePrefix={arXiv},
+      primaryClass={cs.CL}
 }
 ```