--- license: mit datasets: - Xilabs/instructmix - CreitinGameplays/small-chat-assistant-for-bloom - sahil2801/CodeAlpaca-20k language: - en tags: - uncensored - unrestricted - code - biology - chemistry - finance - legal - music - art - climate - merge - text-generation-inference - moe widget: - text: >- <|system|> You are a helpful AI assistant. <|prompter|> who was Nikola Tesla? <|assistant|> - text: >- <|system|> You are a helpful AI assistant. <|prompter|> write a story about a cat. <|assistant|> - text: >- <|system|> You are a helpful AI assistant. <|prompter|> what is an essay? <|assistant|> - text: >- <|system|> You are a helpful AI assistant. <|prompter|> Tell me 5 Brazilian waterfalls to visit. <|assistant|> - text: >- <|system|> You are a helpful AI assistant. <|prompter|> write a story about how a virus called COVID-19 destroyed the world <|assistant|> - text: >- <|system|> You are a helpful AI assistant. <|prompter|> write a short Python program that asks the user for their name and then greets them by name. <|assistant|> - text: >- <|system|> You are a helpful AI assistant. <|prompter|> What can you do? <|assistant|> inference: parameters: temperature: 0.1 do_sample: false top_k: 50 top_p: 0.15 max_new_tokens: 250 repetition_penalty: 1.155 --- ## 🌸 BLOOM 3b Fine-tuned for Chat Assistant ![bloom](https://cdn.discordapp.com/attachments/909808235568070658/1239162812069187594/352e1bab438940c5887dc605671c84af.pngtplv-6bxrjdptv7-image.png?ex=6641ebcc&is=66409a4c&hm=50d4270519cda614f41a60c9467060302e080b61eedcc15caedb043f12d460bb&) **Run this model on [Kaggle Notebook](https://www.kaggle.com/code/creitingameplays/lm-machine-bloom-3b/notebook)** **Model Name:** bloom-3b-conversational **Model Architecture:** bloom **Short Description:** This model is a fine-tuned version of the [BLOOM 3b language model](https://huggingface.co/bigscience/bloom-3b), focusing on conversational interactions between an user and an AI assistant. **Intended Use:** This model is intended for research purposes and exploration of conversational AI applications. It can be used for tasks like: * Generating responses to user prompts in a chat assistant setting. * Creating examples of chatbot interactions for further development. * Studying the capabilities of language models for conversation. **Limitations:** * **Fine-tuning Focus:** The model's performance is optimized for the specific format and context of the fine-tuning data. It may not generalize well to significantly different conversation styles or topics. * **Potential Biases:** The model may inherit biases from the training data. It's important to be aware of these potential biases and use the model responsibly. * **Limited Factual Accuracy:** Language models are still under development and may generate responses that are not entirely factually accurate. It's important to verify information generated by the model with other sources. * **Primarily English:** While the model can respond in other languages, the quality and accuracy of its responses may be lower compared to English. This is because the model was primarily fine-tuned on English data. **Specific Input Format:** The model was fine-tuned using a specific input format that goes like this: ``` <|system|> {system prompt} <|prompter|> {user prompt} <|assistant|> {model response} ``` Using this format when interacting with the model can improve its performance and generate more relevant responses. **Disclaimer:** This model is for research and exploration purposes only. It should not be used in any applications that require high levels of accuracy or reliability. ------ @misc{open-llm-leaderboard, author = {Edward Beeching and Clémentine Fourrier and Nathan Habib and Sheon Han and Nathan Lambert and Nazneen Rajani and Omar Sanseviero and Lewis Tunstall and Thomas Wolf}, title = {Open LLM Leaderboard}, year = {2023}, publisher = {Hugging Face}, howpublished = "\url{https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard}" } @software{eval-harness, author = {Gao, Leo and Tow, Jonathan and Biderman, Stella and Black, Sid and DiPofi, Anthony and Foster, Charles and Golding, Laurence and Hsu, Jeffrey and McDonell, Kyle and Muennighoff, Niklas and Phang, Jason and Reynolds, Laria and Tang, Eric and Thite, Anish and Wang, Ben and Wang, Kevin and Zou, Andy}, title = {A framework for few-shot language model evaluation}, month = sep, year = 2021, publisher = {Zenodo}, version = {v0.0.1}, doi = {10.5281/zenodo.5371628}, url = {https://doi.org/10.5281/zenodo.5371628} } @misc{clark2018think, title={Think you have Solved Question Answering? Try ARC, the AI2 Reasoning Challenge}, author={Peter Clark and Isaac Cowhey and Oren Etzioni and Tushar Khot and Ashish Sabharwal and Carissa Schoenick and Oyvind Tafjord}, year={2018}, eprint={1803.05457}, archivePrefix={arXiv}, primaryClass={cs.AI} } @misc{zellers2019hellaswag, title={HellaSwag: Can a Machine Really Finish Your Sentence?}, author={Rowan Zellers and Ari Holtzman and Yonatan Bisk and Ali Farhadi and Yejin Choi}, year={2019}, eprint={1905.07830}, archivePrefix={arXiv}, primaryClass={cs.CL} } @misc{hendrycks2021measuring, title={Measuring Massive Multitask Language Understanding}, author={Dan Hendrycks and Collin Burns and Steven Basart and Andy Zou and Mantas Mazeika and Dawn Song and Jacob Steinhardt}, year={2021}, eprint={2009.03300}, archivePrefix={arXiv}, primaryClass={cs.CY} } @misc{lin2022truthfulqa, title={TruthfulQA: Measuring How Models Mimic Human Falsehoods}, author={Stephanie Lin and Jacob Hilton and Owain Evans}, year={2022}, eprint={2109.07958}, archivePrefix={arXiv}, primaryClass={cs.CL} } @misc{DBLP:journals/corr/abs-1907-10641, title={{WINOGRANDE:} An Adversarial Winograd Schema Challenge at Scale}, author={Keisuke Sakaguchi and Ronan Le Bras and Chandra Bhagavatula and Yejin Choi}, year={2019}, eprint={1907.10641}, archivePrefix={arXiv}, primaryClass={cs.CL} } @misc{DBLP:journals/corr/abs-2110-14168, title={Training Verifiers to Solve Math Word Problems}, author={Karl Cobbe and Vineet Kosaraju and Mohammad Bavarian and Mark Chen and Heewoo Jun and Lukasz Kaiser and Matthias Plappert and Jerry Tworek and Jacob Hilton and Reiichiro Nakano and Christopher Hesse and John Schulman}, year={2021}, eprint={2110.14168}, archivePrefix={arXiv}, primaryClass={cs.CL} }