Model Card: QwOwO-1.5B

Experimental Status: Highly Experimental.

Base Model: Qwen2.5-1.5B-Instruct

Description: QwOwO-1.5B is a fine-tuned version of Qwen2.5-1.5B-Instruct, trained on a dataset rich in OwO/UwU-style language, internet slang, and generally cutesy expressions. The primary goal was to create a model that generates text in this specific, niche style.

Prompt Format:

  • Compatibility: ChatML/Qwen
  • Best Performance: The model performs best when the prompt includes some elements of the OwO/UwU style. It's primed to continue and amplify this kind of tone. Plain instructions may not always produce the desired level of "OwO-fication."

Strengths:

  • OwO/UwU Style Generation: QwOwO-1.5B can generate text with OwO/UwU characteristics, including "OwO"s.
  • Novelty and Experimentation: This model is intended for fun, experimentation, and exploring the effects of stylistic fine-tuning. The output is often humorous (intentionally or unintentionally).

Limitations & Risks:

  • Inconsistency: The model's ability to consistently maintain the OwO/UwU style is limited. While it produces "OwO"s, the overall quality and stylistic coherence may vary significantly. It's not guaranteed to "OwO-fy" every response perfectly.

  • Reduced Instruction Following: The fine-tuning process has significantly impacted the model's ability to follow instructions reliably. Expect deviations from the prompt and potentially nonsensical outputs. Factual accuracy is not guaranteed.

  • Over-the-top Cuteness (Potentially): The generated text may contain a high density of OwO/UwU language, which can be overwhelming or annoying to some users.

  • Unpredictable Behavior: The model's behavior is unpredictable, especially with complex or unusual prompts. Unexpected, inappropriate, or nonsensical responses are possible.

  • Intended Use:

  • Novelty Applications: Chatbots, interactive fiction, or other applications where a unique and quirky text style is desired.

  • Entertainment and Amusement: Generating humorous or unusual text. Not intended for serious or critical applications.

Ethical Considerations & Safety:

  • Responsible Use is Crucial: Use QwOwO-1.5B responsibly and be aware of its limitations. Avoid using it to generate harmful, offensive, misleading, or unethical content.
  • Output is User Responsibility: You are fully responsible for the text generated by this model and its consequences.
  • No Guarantees: The developers make no guarantees about the model's performance, safety, accuracy, or ethical behavior.

Disclaimer: QwOwO-1.5B is provided "as is," without any warranties. Use it at your own risk. We encourage feedback on your experiences! :3

Thanks again Gemini

Built with Axolotl

Downloads last month
3
Safetensors
Model size
1.54B params
Tensor type
BF16
·
Inference Providers NEW
This model is not currently available via any of the supported Inference Providers.
The model cannot be deployed to the HF Inference API: The model has no library tag.

Model tree for SaisExperiments/QwOwO-1.5B

Base model

Qwen/Qwen2.5-1.5B
Finetuned
(312)
this model
Quantizations
1 model

Dataset used to train SaisExperiments/QwOwO-1.5B

Collection including SaisExperiments/QwOwO-1.5B