---
language:
- en
license: apache-2.0
tags:
- text-generation
- nlp
datasets:
- DeepMath103K
metrics:
- avg@1 / pass@k
base_model:
- Deepseek-R1-Distill-Qwen-1.5B
pipeline_tag: text-generation
library_name: transformers
arxiv: 2604.10688
---

# Model Name

SCOPE-Deepseek-R1-Distill-Qwen-1.5B

This model is introduced in the paper 
[SCOPE: Signal-Calibrated On-Policy Distillation Enhancement with Dual-Path Adaptive Weighting](https://arxiv.org/abs/2604.10688) 
and is developed by the **Longcat Interaction Team**.

## Model Details

### Model Description

- **Developed by:** Longcat Interaction Team
- **Language(s) (NLP):** English
- **License:** Apache 2.0
- **Finetuned from model:** Deepseek-R1-Distill-Qwen-1.5B
- **Paper:** [arxiv.org/abs/2604.10688](https://arxiv.org/abs/2604.10688)

### Model Sources

- **Repository:** [https://github.com/machine981/SCOPE](https://github.com/machine981/SCOPE)
- **Paper:** [https://arxiv.org/abs/2604.10688](https://arxiv.org/abs/2604.10688)

## Uses

### Direct Use

This model can be used directly for text generation (like MATH reasoning) without any additional fine-tuning.

## How to Get Started with the Model

Use the code below to get started with the model:

```python
from transformers import AutoTokenizer, AutoModelForCausalLM  # adjust as needed

tokenizer = AutoTokenizer.from_pretrained("Machine981/SCOPE-Deepseek-R1-Distill-Qwen-1.5B")
model = AutoModelForCausalLM.from_pretrained("Machine981/SCOPE-Deepseek-R1-Distill-Qwen-1.5B")

inputs = tokenizer("Your input text here", return_tensors="pt")
outputs = model.generate(**inputs)
print(tokenizer.decode(outputs[0], skip_special_tokens=True))