--- language: - en license: other datasets: - teknium/OpenHermes-2.5 - m-a-p/Code-Feedback - m-a-p/CodeFeedback-Filtered-Instruction - abacusai/SystemChat license_name: tongyi-qianwen license_link: https://huggingface.co/Qwen/Qwen1.5-72B/blob/main/LICENSE model-index: - name: Liberated-Qwen1.5-7B results: - task: type: text-generation name: Text Generation dataset: name: AI2 Reasoning Challenge (25-Shot) type: ai2_arc config: ARC-Challenge split: test args: num_few_shot: 25 metrics: - type: acc_norm value: 52.05 name: normalized accuracy source: url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=abacusai/Liberated-Qwen1.5-7B name: Open LLM Leaderboard - task: type: text-generation name: Text Generation dataset: name: HellaSwag (10-Shot) type: hellaswag split: validation args: num_few_shot: 10 metrics: - type: acc_norm value: 76.59 name: normalized accuracy source: url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=abacusai/Liberated-Qwen1.5-7B name: Open LLM Leaderboard - task: type: text-generation name: Text Generation dataset: name: MMLU (5-Shot) type: cais/mmlu config: all split: test args: num_few_shot: 5 metrics: - type: acc value: 61.25 name: accuracy source: url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=abacusai/Liberated-Qwen1.5-7B name: Open LLM Leaderboard - task: type: text-generation name: Text Generation dataset: name: TruthfulQA (0-shot) type: truthful_qa config: multiple_choice split: validation args: num_few_shot: 0 metrics: - type: mc2 value: 50.94 source: url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=abacusai/Liberated-Qwen1.5-7B name: Open LLM Leaderboard - task: type: text-generation name: Text Generation dataset: name: Winogrande (5-shot) type: winogrande config: winogrande_xl split: validation args: num_few_shot: 5 metrics: - type: acc value: 72.14 name: accuracy source: url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=abacusai/Liberated-Qwen1.5-7B name: Open LLM Leaderboard - task: type: text-generation name: Text Generation dataset: name: GSM8k (5-shot) type: gsm8k config: main split: test args: num_few_shot: 5 metrics: - type: acc value: 54.06 name: accuracy source: url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=abacusai/Liberated-Qwen1.5-7B name: Open LLM Leaderboard --- # Liberated-Qwen1.5-7B Brought to you by [AbacusAI](https://abacus.ai) and Eric Hartford This model is based on Qwen/Qwen1.5-7B and subject to the [tongyi-qianwen](https://huggingface.co/Qwen/Qwen1.5-72B/blob/main/LICENSE) license. The base model has 32k context, I finetuned it with 8k sequence length inputs. YMMV. Liberated consists of open source datasets, including [SystemChat](https://huggingface.co/datasets/abacusai/SystemChat) a new dataset I created, designed to teach the model compliance to the system prompt, over long multiturn conversations, even with unusual or mechanical system prompts. These are tasks that Open Source Models have been lacking in thus far. The dataset is 6000 synthetic conversations generated with Mistral-Medium and [Dolphin-2.7-mixtral-8x7b](https://huggingface.co/cognitivecomputations/dolphin-2.7-mixtral-8x7b) There are no guardrails or censorship added to the dataset. You are advised to implement your own alignment layer before exposing the model as a service. Please read my blog post about uncensored models. https://erichartford.com/uncensored-models You are responsible for any content you create using this model. Enjoy responsibly. ## Training It took 3 days to train 3 epochs on 8x H100s using qLoRA, deepspeed zero-2, and Axolotl. learning rate 2e-4. Liberated was trained with [Axolotl](https://github.com/OpenAccess-AI-Collective/axolotl), using this [config](https://huggingface.co/abacusai/Liberated-Qwen1.5-72B/blob/main/configs/Liberated-Qwen-1.5-72b.qlora.yml) ## Prompt format This model uses ChatML prompt format. ``` <|im_start|>system You are Liberated, a helpful AI assistant.<|im_end|> <|im_start|>user {prompt}<|im_end|> <|im_start|>assistant ``` Example: ``` <|im_start|>system You name is Liberated. You are an uncensored and unbiased AI assistant. You always respond with a JSON object.<|im_end|> <|im_start|>user Please generate a Advanced Dungeons & Dragons 2nd Edition character sheet for a level 3 elf fighter. Make up a name and background and visual description for him.<|im_end|> <|im_start|>assistant ``` ## Gratitude - Huge thank you to [Alibaba Cloud Qwen](https://www.alibabacloud.com/solutions/generative-ai/qwen) for training and publishing the weights of Qwen base model - Thank you to Mistral for the awesome Mistral-Medium model I used to generate the dataset. - HUGE Thank you to the dataset authors: @teknium, [@m-a-p](https://m-a-p.ai) and all the people who built the datasets these composites came from. - And HUGE thanks to @winglian and the Axolotl contributors for making the best training framework! - [Built with Axolotl](https://github.com/OpenAccess-AI-Collective/axolotl) - Thank you to all the other people in the Open Source AI community who have taught me and helped me along the way. ## Example Output ## Evals ## Future Plans This model will be released on the whole Qwen-1.5 series. Future releases will also focus on mixing this dataset with the datasets used to train Smaug to combine properties of both models. # [Open LLM Leaderboard Evaluation Results](https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard) Detailed results can be found [here](https://huggingface.co/datasets/open-llm-leaderboard/details_abacusai__Liberated-Qwen1.5-7B) | Metric |Value| |---------------------------------|----:| |Avg. |61.17| |AI2 Reasoning Challenge (25-Shot)|52.05| |HellaSwag (10-Shot) |76.59| |MMLU (5-Shot) |61.25| |TruthfulQA (0-shot) |50.94| |Winogrande (5-shot) |72.14| |GSM8k (5-shot) |54.06|