---
license: mit
language:
- en
base_model:
- Qwen/Qwen3-8B
tags:
- social-reasoning
- Theory of Mind
- reinforcement-learning
- GRPO
- SIP
datasets:
- Jincenzi/ToMBench_Hard
pipeline_tag: text-generation
---

# SocialR1-8B

**SocialR1-8B** is a social reasoning model built on [Qwen3-8B](https://huggingface.co/Qwen/Qwen3-8B), trained with trajectory-level reinforcement learning (GRPO) using the **Social-R1** framework. It enhances Theory-of-Mind (ToM) and social inference capabilities by aligning reasoning processes with the Social Information Processing (SIP) theory.

📄 **Paper**: [Social-R1: Enhancing Social Reasoning in LLMs through Trajectory-Level Reinforcement Learning](https://arxiv.org/abs/2603.09249) 

## Highlights

- 🧠 **SIP-Guided Reasoning**: Enforces stage-consistent social inference — Cue Encoding → Cue Interpretation → Goal Clarification → Response Generation
- 🎯 **Multi-Dimensional Reward**: Combines structural reward, content reward, inference efficiency, and format reward with curriculum-style weighting
- 📊 **Strong Performance**: Matches or outperforms substantially larger baselines across static MCQ benchmarks, open-ended generation (FanToM), and interactive settings (SOTOPIA)

## Usage

```python
from transformers import AutoModelForCausalLM, AutoTokenizer

model_name = "Jincenzi/SocialR1-8B"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(model_name, torch_dtype="auto", device_map="auto")

messages = [
    {"role": "user", "content": "You should first think about the reasoning process in the mind and then provide with the answer.The reasoning process and answer are enclosed within <think> </think> and <Answer> </Answer> tags, respectively."}
]
text = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
inputs = tokenizer([text], return_tensors="pt").to(model.device)
outputs = model.generate(**inputs, max_new_tokens=2048)
print(tokenizer.decode(outputs[0], skip_special_tokens=True))
```

## Training Details

- **Base Model**: Qwen3-8B
- **Training Method**: Group Relative Policy Optimization (GRPO)
- **Training Steps**: 600
- **Hardware**: 8× NVIDIA A100 (80GB)
- **Reward Design**: SIP structural reward ($R_\text{struct}$) + SIP content reward ($R_\text{cont}$) + inference efficiency ($R_\text{len}$) + format reward ($R_\text{fmt}$)

## Evaluation

SocialR1-8B is evaluated across three complementary settings:

- **Static MCQ**: ToMBench, ToMBench-Hard, SocialIQA, SimpleToM, EmoBench, MotiveBench, Hi-ToM, TactfulToM
- **Open-ended Generation**: FanToM
- **Interactive Social Intelligence**: SOTOPIA

## Related Resources

| Resource | Link |
|----------|------|
| Paper | [arXiv:2603.09249](https://arxiv.org/abs/2603.09249) |
| SocialR1-4B | [Jincenzi/SocialR1-4B](https://huggingface.co/Jincenzi/SocialR1-4B) |

## Citation

```BibTeX
@inproceedings{wu2026socialr1,
  title={Social-R1: Enhancing Social Reasoning in LLMs through Trajectory-Level Reinforcement Learning},
  author={Wu, Jincenzi and Lei, Yuxuan and Lian, Jianxun and Huang, Yitian and Zhou, Lexin and Li, Haotian and Yang, Deng and Xie, Xing and Meng, Helen},
  booktitle={Arxiv},
  year={2026}
}
```

## Contact

For questions or discussions, please contact [jincenziwu@gmail.com](mailto:jincenziwu@gmail.com).