Data

Tools

Indexes

Applications

Experiments

Awesome

An open API service indexing awesome lists of open source software.

https://github.com/cxcscmu/AutoGEO

AutoGEO: a framework to automatically learn generative engine preferences, and rewrite web contents for more traction.
https://github.com/cxcscmu/AutoGEO

ai-search-engine ai-search-optimization content-optimization generative-ai generative-engine-optimization generative-search grpo large-language-models retrieval-augmented-generation search visibility

Last synced: 3 months ago
JSON representation

AutoGEO: a framework to automatically learn generative engine preferences, and rewrite web contents for more traction.

Awesome Lists containing this project

README 

          
# AutoGEO



[Project Page](https://zhongshsh.github.io/AutoGEO/) | [Paper](https://arxiv.org/abs/2510.11438) | [Demo](https://huggingface.co/spaces/cx-cmu/AutoGEO_Mini) 



**AutoGEO** is a framework for Automatic Generative Engine Optimization (GEO) that helps web content gain higher visibility in LLM-generated answers.



📄 **Paper:** "What Generative Search Engines Like and How to Optimize Web Content Cooperatively"  

👥 **Authors:** Yujiang Wu*, Shanshan Zhong*, Yubin Kim, Chenyan Xiong (*Equal contribution)



## 🔍 Overview



AutoGEO automatically extracts content preference rules from generative engines and rewrites documents to maximize visibility while preserving accuracy.



**How GEO models work:**

- **Input:** Target document

- **Output:** Rewritten document with higher visibility in generative engine (GE) responses

- **Goal:** Maximize visibility without harming GE utility



**Three core components:**



1. **Rule Extraction** — Automatically mines content preferences from GEs.

2. **AutoGEOAPI** — Prompt-based GEO model using extracted rules

3. **AutoGEOMini** — Cost-effective GEO model trained with reinforcement learning



**Evaluation metrics:** **GEO score** (visibility) and **GEU score** (utility)



## News



- 🔥 **[2026-01-17]**: We have released our [AutoGEOMini Demo](https://huggingface.co/spaces/cx-cmu/AutoGEO_Mini). Feel free to try it out!

- 🔥 **[2026-01-17]**: We have released our checkpoints ([E-commerce](https://huggingface.co/cx-cmu/AutoGEO_mini_Qwen1.7B_Ecommerce), [GEO-Bench](https://huggingface.co/cx-cmu/AutoGEO_mini_Qwen1.7B_GEOBench), [Researchy-GEO](https://huggingface.co/cx-cmu/AutoGEO_mini_Qwen1.7B_ResearchyGEO)). 

- 🔥 **[2025-12-08]**: We have released our code and datasets ([E-commerce](https://huggingface.co/datasets/cx-cmu/E-commerce), [GEO-Bench](https://huggingface.co/datasets/cx-cmu/GEO-Bench), [Researchy-GEO](https://huggingface.co/datasets/cx-cmu/Researchy-GEO)). 

- 🔥 **[2025-10-11]**: Our paper is now available on [arXiv](https://arxiv.org/pdf/2510.11438). Check it out!



## 🚀 Installation



For using AutoGEOAPI and rule extraction:



```bash

# Clone the repository

git clone --recursive https://github.com/cxcscmu/AutoGEO

cd AutoGEO



# Run installation script

bash install.sh



# Activate environment

conda activate autogeo



# Configure API keys (required)

nano keys.env  # Add your API keys

```



Optional: For training AutoGEOMini models:



```bash

# First complete Option 1, then:

conda activate autogeo

bash install_mini.sh

```



**⚠️ Note:** AutoGEOMini requires:

- CUDA-compatible GPU * 2 (A100 40GB+ recommended)

- ~4h for SFT and ~48h for GRPO on Researchy-GEO



## ⚡ Quick Start



Rewrite a document using AutoGEOAPI:



```python

from autogeo.rewriters import rewrite_document



rewritten_text = rewrite_document(

    document="AutoGEO automatically extracts content preference rules from generative engines and rewrites documents to maximize visibility while preserving accuracy.",

    dataset="Researchy-GEO",   # Options: E-commerce, GEO-Bench, Researchy-GEO

    engine_llm="gemini"        # Options: gemini, gpt, claude

)



print(rewritten_text)

```



## 🧩 Rule Extraction



Extract content preference rules from a generative engine (example: Gemini on E-commerce):



```bash

python -m autogeo.extract_rules \

    --dataset E-commerce \

    --engine_llm gemini-2.5-flash-lite

```



Rules are saved to: `data/E-commerce/rule_sets/gemini-2.5-flash-lite/`. 

**Tips:**

- Reduce concurrency if hitting API rate limits: `--max_workers 4`

- Test on a small subset: `--num_examples 10`



Use extracted or custom rules for rewriting:



```python

from autogeo.rewriters import rewrite_document



rewritten_text = rewrite_document(

    document="Your document text here",

    rule_path=f"data/{dataset}/rule_sets/{engine_llm}/merged_rules.json"

)

```



**Custom rules format:** JSON file with root key `"filtered_rules"`



## 🧩 AutoGEOAPI



AutoGEO provides a unified evaluation framework for all models.



**Model types:**

- `vanilla` — Original documents (baseline)

- `autogeo_api` — Rewritten documents generated by prompt-based GEO model

- `autogeo_mini` — Rewritten documents generated by cost-effective GEO model



**Evaluate baseline:**

```bash

python -m autogeo.evaluate \

    --model vanilla \

    --dataset E-commerce \

    --engine_llm gemini-2.5-flash-lite

```



**Evaluate AutoGEOAPI:**

```bash

python -m autogeo.evaluate \

    --model autogeo_api \

    --dataset E-commerce \

    --engine_llm gemini-2.5-flash-lite

```



**Tips:**

- Include GEU score: `--need_geu_score`

- Test subset: `--num_examples 10`



## 🧩 AutoGEOMini



Train a cost-effective GEO model using reinforcement learning.



**Step 1: Cold Start (Supervised Fine-Tuning)**

```bash

bash run_cold_start.sh E-commerce

```

Using training data (`data/E-commerce/RL/finetune.json`) and starts LLaMA-Factory training. Checkpoint saved to `outputs/E-commerce/cold_start`.



**Step 2: GRPO Training**

```bash

bash run_grpo.sh E-commerce

```

Trains the model using Group Relative Policy Optimization. Checkpoint saved to `outputs/E-commerce/grpo`.



If you encounter GRPO-related dependency errors, it is usually caused by version conflicts between LLaMA-Factory and open-r1. To resolve this, reinstall open-r1:

```

cd open-r1

GIT_LFS_SKIP_SMUDGE=1 pip install -e ".[dev]"

```



**Step 3: Evaluation**

```bash

python -m autogeo.evaluate \

    --model autogeo_mini \

    --model_path outputs/E-commerce/grpo \

    --dataset E-commerce \

    --engine_llm gemini-2.5-flash-lite

```



## 📚 Supported Datasets & Engines & Metrics



**Datasets:**

- **Researchy-GEO** — Academic dataset

- **E-commerce** — Commercial dataset

- **GEO-Bench** — Benchmark from [GEO](https://generative-engines.com/GEO/)



**Generative Engines:**

- **Gemini** (e.g., `gemini-2.5-flash-lite`)

- **GPT** (e.g., `gpt-4o-mini`)

- **Claude** (e.g., `claude-3-5-sonnet-20241022`)



**Metrics:**

- **GEO Score** — Visibility (position, token count, citation frequency)

- **GEU Score** — Utility (citation quality, keypoint coverage, response quality)



## 🙏 Acknowledgements



We thank the authors of [GEO](https://generative-engines.com/GEO/), [AutoRule](https://github.com/cxcscmu/AutoRule), [LLaMA-Factory](https://github.com/hiyouga/LLaMA-Factory), [open-r1](https://github.com/huggingface/open-r1), and [DeepResearchGym](https://github.com/cxcscmu/deepresearch_benchmarking) for their inspiring works. We also thank [Qwen3](https://github.com/QwenLM/Qwen3) and [DeepSeek-R1](https://huggingface.co/deepseek-ai/DeepSeek-R1) for their excellent models.



## 📖 Citation



If you find AutoGEO useful, please cite:



```bibtex

@article{wu2025generative,

  title={What Generative Search Engines Like and How to Optimize Web Content Cooperatively},

  author={Wu, Yujiang and Zhong, Shanshan and Kim, Yubin and Xiong, Chenyan},

  journal={arXiv preprint arXiv:2510.11438},

  year={2025}

}

```