---
license: cc-by-nc-nd-4.0
---

# REDDME

This repository contains the dataset introduced in the paper:


- Paper: [Assess and Prompt: A Generative RL Framework for Improving Engagement in Online Mental Health Communities](https://arxiv.org/abs/2508.16788)
- Published in **Findings of EMNLP 2025**  

## 📖 Description

We propose **REDDME**, a manually annotated corpus extended from the publicly available mental health subreddit corpus, **BeCOPE** (Srivastava et al., 2025).  

- **BeCOPE** categorizes posts into three categories:  
  - **Interactive** → back-and-forth conversations between the OP (original poster) and peers  
  - **Non-interactive** → post engages peers, but the OP does not reply  
  - **Isolated** → post receives no comments  

- **REDDME** selects **4,760 posts** from BeCOPE and **manually annotates** them with **support attributes**:
  - **Event**  
  - **Effect**  
  - **Requirement**  

Each attribute is annotated with **spans (rationales)** and an **intensity level**.

## Citation

If you use this work, please cite:

```bibtex
@misc{gaur2025assesspromptgenerativerl,
      title={Assess and Prompt: A Generative RL Framework for Improving Engagement in Online Mental Health Communities}, 
      author={Bhagesh Gaur and Karan Gupta and Aseem Srivastava and Manish Gupta and Md Shad Akhtar},
      year={2025},
      eprint={2508.16788},
      archivePrefix={arXiv},
      primaryClass={cs.CL},
      url={https://arxiv.org/abs/2508.16788}, 
}