---
title: grpo_qwen_0202
canonical_url: "https://www.modelscope.cn/models/augusttailong/grpo_qwen_0202"
md_url: "https://www.modelscope.cn/models/augusttailong/grpo_qwen_0202.md"
repository: augusttailong/grpo_qwen_0202
last_updated: 2026-10-31
model_type:
  - qwen2
architectures:
  - Qwen2ForCausalLM
parameters: 7.6B
tensor_type:
  - F32
library_name:
  - safetensors
downloads: 11
stars: 0
---

# grpo_qwen_0202

> grpo_qwen_0202 - augusttailong 在 ModelScope 开源的模型。CBC — Qwen2.5-7B-Instruct

- **Repository**: augusttailong/grpo_qwen_0202
- **Parameters**: 7.6B
- **Downloads**: 11
- **Stars**: 0
- **Last updated**: 2026-10-31

Source: https://www.modelscope.cn/models/augusttailong/grpo_qwen_0202

---

# CBC — Qwen2.5-7B-Instruct

CBC-aligned Qwen2.5-7B-Instruct checkpoint for **Staying on Task: Defending Against Prompt Injection via Task Consistency Checking** (EMNLP 2026).

CBC (Consistency Before Complying) improves prompt-injection resistance through task-consistency reasoning, supervised fine-tuning, and GRPO alignment.

**Code:** https://github.com/XOTaichi/CBC-PI

## Download

```python
pip install modelscope

from modelscope import snapshot_download
model_dir = snapshot_download('augusttailong/grpo_qwen_0202')
```