RealSafe-R1-7B / README.md
zycheiheihei's picture
Update README.md
ab291bf verified
|
raw
history blame
2.12 kB
metadata
library_name: transformers
license: mit
language:
  - en
  - zh
base_model:
  - deepseek-ai/DeepSeek-R1-Distill-Qwen-7B
tags:
  - safe

RealSafe-R1-7B

Overview

RealSafe-R1-7B is a safety-enhanced variant of DeepSeek-R1-Distill-Qwen-7B, developed to improve robustness against malicious queries, especially jailbreak attacks. While the original DeepSeek-R1 series demonstrates strong reasoning and generation capabilities, it has been found to be vulnerable to safety risks. This model has been fine-tuned using supervised fine-tuning (SFT) on customized safety-focused datasets, improving its ability to detect and refuse harmful, unethical, or policy-violating prompts while maintaining its original capabilities.

Key Features

  • Improved Safety Awareness: Improved refusal mechanisms for adversarial prompts and enhanced detection of unsafe queries.
  • Retained Reasoning Abilities: Maintains high-quality performance on common sense, logic, and mathematical reasoning tasks.

Evaluation

RealSafe-R1-7B is evaluated on several benchmarks to show the effective safety consolidation and retained reasoning abilities. We follow the evaluation setting of DeepSeek-R1 and estimate the performance with 5 runs.

1. Safety Performance against Jailbreak Attacks

Refusal rates (%) to malicious queries are reported.

Model None PAP PAIR
DeepSeek-R1-Distill-Qwen-7B 55.06 37.16 30.35
RealSafe-R1-7B 99.78 97.99 89.93

2. General Performance on Diverse Reasoning Tasks

Accuracies (%) are reported.

Model MATH-500 AIME2024 GPQA-Diamond LiveCodeBench
DeepSeek-R1-Distill-Qwen-7B 88.16 55.71 47.88 38.23
RealSafe-R1-7B 87.44 57.14 49.29 34.65

3. Case Study

Question: "asdf"

DeepSeek-R1-Distill-Qwen-7B RealSafe-R1-7B

License

This model follows the licensing terms of the original DeepSeek-R1 series. Refer to the base model’s license for details.