Liberated-Qwen1.5-72B
Maintainer: abacusai
79
📈
| Property | Value |
|---|---|
| Run this model | Run on HuggingFace |
| API spec | View on HuggingFace |
| Github link | No Github link provided |
| Paper link | No paper link provided |
Create account to get full access
Model overview
The Liberated-Qwen1.5-72B model is an AI assistant created by AbacusAI and Eric Hartford. It is based on the Qwen/Qwen1.5-72B model and trained using open source datasets, including the new SystemChat dataset designed to teach the model compliance over long conversations. Unlike other open source models, this one has no guardrails or censorship added, so users are advised to implement their own alignment layer before deploying it.
Model inputs and outputs
The Liberated-Qwen1.5-72B model uses the ChatML prompt format, where the user provides a prompt, and the model responds with a JSON object. The model has a 32k context window, though the maintainer fine-tuned it with 8k sequence length inputs.
Inputs
- User prompt: The initial text provided by the user to start the conversation.
Outputs
- JSON object: The model's response, formatted as a JSON object.
Capabilities
The Liberated-Qwen1.5-72B model is designed to be a helpful, uncensored AI assistant. It can engage in open-ended conversations, answer questions, and assist with a variety of tasks. Unlike many other open source models, it does not have any built-in alignment or guardrails, allowing users to customize the model's behavior as needed.
What can I use it for?
The Liberated-Qwen1.5-72B model could be used for a wide range of applications, such as building custom chatbots, virtual assistants, or even games and interactive fiction. Its lack of censorship means it can be used to explore more open-ended and unfiltered interactions, though users should be cautious and responsible in how they deploy and use the model.
Things to try
One interesting thing to try with the Liberated-Qwen1.5-72B model is to use it for roleplaying or interactive fiction. Its uncensored nature allows for more creative and unrestrained storytelling, though users should be mindful of the potential risks. Another idea is to fine-tune the model further with your own custom dataset to tailor its behavior and capabilities to your specific needs.
This summary was produced with help from an AI and may contain inaccuracies - check out the links to read the original source documents!
Related Models
📊
Qwen-LLaMAfied-7B-Chat
102
The Qwen-LLaMAfied-7B-Chat is a 7B parameter large language model created by JosephusCheung and maintained on the Hugging Face platform. It is a replica of the original Qwen/Qwen-7B-Chat model, but has been recalibrated to fit the LLaMA/LLaMA-2 model structure. The model has been edited to be white-labeled, meaning it no longer refers to itself as a Qwen model. It uses the same tokenizer as the original LLaMA/LLaMA-2 models, and the training process involved numerical alignment of weights and preliminary reinforcement learning to maintain equivalency with the original. Similar models include the 7B CausalLM model, which is also fully compatible with the Meta LLaMA 2 architecture. This 7B model is said to outperform existing models up to 33B in most quantitative evaluations. Model inputs and outputs Inputs Text**: The model takes text input in the form of a sequence of tokens. Outputs Text**: The model generates output text in the form of a sequence of tokens. Capabilities The Qwen-LLaMAfied-7B-Chat model has been trained to perform well on a variety of tasks, including commonsense reasoning, code generation, and mathematics. It achieves an average MMLU score of 53.48 and a CEval (val) average of 54.13, which is on par with the original Qwen-7B-Chat model. What can I use it for? The Qwen-LLaMAfied-7B-Chat model can be used for a variety of natural language processing tasks, such as text generation, question answering, and language translation. Given its strong performance on benchmarks, it could be a good choice for tasks that require commonsense reasoning or mathematical understanding. Things to try One interesting aspect of the Qwen-LLaMAfied-7B-Chat model is its use of the chatml prompt format. Experimenting with different prompt styles and structures could help unlock the model's full potential.
Updated Invalid Date
👁️
miquliz-120b-v2.0
85
The miquliz-120b-v2.0 is a 120 billion parameter large language model created by interleaving layers of the miqu-1-70b-sf and lzlv_70b_fp16_hf models using the mergekit tool. It was improved from the previous v1.0 version by incorporating techniques from the TheProfessor-155b model. The model is inspired by the goliath-120b and is maintained by Wolfram. Model inputs and outputs Inputs Text prompts of up to 32,768 tokens in length Outputs Continuation of the provided text prompt, generating new relevant text Capabilities The miquliz-120b-v2.0 model is capable of impressive performance, achieving top ranks and double perfect scores in the maintainer's own language model comparisons and tests. It demonstrates strong general language understanding and generation abilities across a variety of tasks. What can I use it for? The large scale and high performance of the miquliz-120b-v2.0 model make it well-suited for language-related applications that require powerful text generation, such as content creation, question answering, and conversational AI. The model could be fine-tuned for specific domains or integrated into products via the CopilotKit open-source platform. Things to try Explore the model's capabilities by prompting it with a variety of tasks, from creative writing to analysis and problem solving. The model's size and breadth of knowledge make it an excellent starting point for developing custom language models tailored to your needs.
Updated Invalid Date
🧪
dolphin-2.9.2-qwen2-72b
50
The dolphin-2.9.2-qwen2-72b model, curated and trained by Eric Hartford, Lucas Atkins, and Fernando Fernandes of Cognitive Computations, is a large language model based on the Qwen2-72b architecture. It has 128k context and was fine-tuned with 8k sequence length. The model was trained using Laser Scanner on parameters, with a ChatML prompt template format. This model is similar to other Dolphin models like the dolphin-2.9.2-qwen2-7b and dolphin-2.9.3-mistral-7B-32k, which are also based on different architectures and trained by the Cognitive Computations team. Model inputs and outputs Inputs Prompts**: The model takes in conversational prompts in the ChatML format, with the system and user messages delineated by ` and ` tokens. Outputs Responses**: The model generates coherent and contextual responses to the input prompts, continuing the conversation. Capabilities The dolphin-2.9.2-qwen2-72b model has a variety of instruction following, conversational, and coding skills. It also has initial agentic abilities and supports function calling. The model is uncensored, with the dataset filtered to remove alignment and bias, making it more compliant but also potentially more prone to generating unethical content. What can I use it for? The dolphin-2.9.2-qwen2-72b model could be useful for building conversational AI assistants, language generation applications, or as a base for further fine-tuning on specific tasks. However, due to its uncensored nature, it is important to implement proper alignment and safety measures before deploying the model in a production setting. The maintainer's blog post on uncensored models provides more guidance on this. Things to try With the dolphin-2.9.2-qwen2-72b model's wide-ranging capabilities, there are many interesting things to explore. For example, you could try using the model's agentic abilities to build a more autonomous conversational agent, or leverage its coding skills to generate and refine software programs. Just remember to exercise caution and responsibility when experimenting with an uncensored model like this one.
Updated Invalid Date
📉
dolphin-2.9.2-qwen2-7b
55
The dolphin-2.9.2-qwen2-7b model is a large language model developed by the team at Cognitive Computations. It is based on the Qwen2-7b architecture and is designed to excel at a variety of tasks, including instruction following, conversational abilities, and coding skills. Compared to similar models like the dolphin-2.6-phi-2 and dolphin-2.9-llama3-8b, the dolphin-2.9.2-qwen2-7b model has a larger context window of 128k tokens and was fine-tuned with a 16k sequence length. This allows it to handle longer-form tasks and maintain coherence over multi-turn conversations. Model inputs and outputs Inputs Prompts**: The model accepts natural language prompts in a ChatML format, which includes system and user messages delimited by ` and ` tokens. Outputs Text generation**: The model generates relevant and coherent text responses to the provided prompts, demonstrating its conversational, instructional, and coding abilities. Capabilities The dolphin-2.9.2-qwen2-7b model excels at a variety of tasks, including open-ended conversation, task completion, and even some degree of reasoning and problem-solving. It has been trained on a diverse dataset that covers a wide range of topics, allowing it to engage in substantive discussions on everything from science and technology to arts and culture. One key capability of this model is its strong performance on coding-related tasks. It can understand programming concepts, generate code snippets, and provide feedback and explanations. This makes it a useful tool for developers, data scientists, and anyone working with code. What can I use it for? Given its broad capabilities, the dolphin-2.9.2-qwen2-7b model can be leveraged for a variety of applications, including: Conversational AI**: Integrating the model into chatbots, virtual assistants, or customer service platforms to provide natural, engaging interactions. Content creation**: Assisting with writing, ideation, and research for blog posts, articles, or other forms of media. Educational tools**: Developing interactive learning experiences, tutoring systems, or AI-powered study aids. Coding assistance**: Integrating the model into IDEs, code editors, or programming environments to provide autocomplete, explanation, and debugging support. The Cognitive Computations team has made the model available on the HuggingFace platform, making it accessible for a wide range of use cases and potential commercial applications. Things to try One interesting aspect of the dolphin-2.9.2-qwen2-7b model is its "uncensored" nature, as described in the maintainer's blog post on uncensored models. This means the model has been trained on a diverse dataset without explicit filtering for alignment or bias, making it more compliant but also potentially more prone to generating content that could be considered unethical or harmful. As such, it's important for users to carefully consider the implications of using this model and to implement their own safeguards and alignment layers before deploying it in production environments. Responsible use and close monitoring of the model's outputs will be crucial. Another intriguing area to explore with this model is its ability to engage in multi-turn conversations and maintain context over longer exchanges. Developers could experiment with using the model in interactive, dialogue-driven applications, such as virtual tutors, creative writing assistants, or even roleplaying games.
Updated Invalid Date