|
|
--- |
|
|
model-index: |
|
|
- name: Ursa_Minor0.4 |
|
|
model-id: Sculptor-AI/Ursa_Minor0.4 |
|
|
results: [] |
|
|
--- |
|
|
|
|
|
# Ursa_Minor0.4 |
|
|
|
|
|
## Model Description |
|
|
|
|
|
Ursa_Minor0.4 is a reasoning-focused language model developed by ExplodingCB2 (Sculptor-AI) and hosted on Hugging Face. It is designed to tackle complex reasoning tasks, demonstrating capabilities in multi-step inference, logical deduction, and contextual understanding. |
|
|
|
|
|
**Key Features:** |
|
|
|
|
|
* **Reasoning Prowess:** Emphasizes strong reasoning abilities over sheer memorization, aiming for accurate and logical responses. |
|
|
* **Multi-Step Inference:** Capable of breaking down complex problems into smaller, manageable steps. |
|
|
* **Logical Deduction:** Demonstrates proficiency in applying logical rules and principles to arrive at valid conclusions. |
|
|
* **Contextual Understanding:** Exhibits an ability to grasp and utilize contextual information to enhance reasoning accuracy. |
|
|
* **Developed by ExplodingCB2 & Kaileh57 (Sculptor-AI):** A model born from focused research and development in the field of AI reasoning. |
|
|
|
|
|
## Intended Uses |
|
|
|
|
|
* Answering complex questions that require multi-step reasoning. |
|
|
* Solving logical puzzles and problems. |
|
|
* Assisting in tasks that demand contextual understanding and inference. |
|
|
* Research and development in the field of AI reasoning. |
|
|
* Experimentation with advanced prompting techniques. |
|
|
|
|
|
## How to Use |
|
|
|
|
|
You can use the Ursa_Minor0.4 model through the Hugging Face Transformers library. Here's a basic example: |
|
|
|
|
|
```python |
|
|
from transformers import AutoTokenizer, AutoModelForCausalLM |
|
|
|
|
|
tokenizer = AutoTokenizer.from_pretrained("Sculptor-AI/Ursa_Minor") |
|
|
model = AutoModelForCausalLM.from_pretrained("Sculptor-AI/Ursa_Minor") |
|
|
|
|
|
prompt = "What are the prime factors of 42?" |
|
|
inputs = tokenizer(prompt, return_tensors="pt") |
|
|
outputs = model.generate(**inputs, max_new_tokens=50) |
|
|
response = tokenizer.decode(outputs[0], skip_special_tokens=True) |
|
|
|
|
|
print(response) |
