e5-small-lora-ai-generated-detector open-source model - Accurately distinguish between AI and human-written texts

E5 Small Lora Ai Generated Detector

Developed by MayZhou

A text classification model fine-tuned using LoRA technology based on the e5-small model, specifically designed for high-accuracy differentiation between AI-generated and human-written text

Text Classification

Transformers

Open Source License:MIT #AI-generated text detection #LoRA fine-tuning #High-precision classification

Downloads 1,039

Release Time : 11/8/2024

Model Overview

This model employs LoRA (Low-Rank Adaptation) technology to fine-tune the e5-small base model, primarily used for detecting whether text content is AI-generated. While maintaining the lightweight nature of the base model, it significantly improves the recognition accuracy of AI-generated text.

Model Features

Efficient LoRA Fine-tuning

Utilizes low-rank adaptation technology for efficient parameter fine-tuning, significantly reducing training costs while maintaining base model performance

High-precision Detection

Achieves 93.9% accuracy on the RAID test set, significantly outperforming the original base model

Strong Robustness

Exhibits good adaptability to image artifacts and text variations, suitable for various application scenarios

Model Capabilities

AI-generated text detection

Human-written text recognition

Text content classification

Use Cases

Academic Integrity

Plagiarism Detection

Identify potentially AI-generated content in academic papers

Accurately distinguishes between AI-generated and human-written academic texts on test sets

Content Moderation

Social Media Content Review

Automatically identify AI-generated content on social media platforms

Effectively detects tweets generated by models like GPT-4

🚀 My LoRA Fine-Tuned AI-generated Detector

This is an e5-small model fine-tuned with LoRA for sequence classification tasks. It can accurately classify text as either AI-generated or human-written.

Label_0: Represents human-written content.
Label_1: Represents AI-generated content.

📚 Documentation

Model Details

Property	Details
Base Model	`intfloat/e5-small`
Fine-Tuning Technique	LoRA (Low-Rank Adaptation)
Task	Sequence Classification
Use Cases	Text classification for AI-generated detection
Hyperparameters	Learning rate: `5e-5` Epochs: `3` LoRA rank: `8` LoRA alpha: `16`

Training Details

Dataset:
- 10,000 twitters and 10,000 rewritten twitters with GPT-4o-mini.
- 80,000 human-written text from RAID-train.
- 128,000 AI-generated text from RAID-train.
Hardware: Fine-tuned on a single NVIDIA A100 GPU.
Training Time: Approximately 2 hours.
Evaluation Metrics: | Metric | (Raw) E5-small | Fine-tuned | |--------|---------------:|-----------:| | Accuracy | 0.652 | 0.890 | | F1 Score | 0.653 | 0.887 | | AUC | 0.697 | 0.976 |

Collaborators

Menglin Zhou
Jiaping Liu
Xiaotian Zhan

Model Index

Name: A Shared Benchmark for Robust Evaluation of Machine-Generated Text Detectors
- Results:
  - Task:
    - Type: text-classification
  - Dataset:
    - Name: RAID-test
    - Type: RAID-test
  - Metrics:
    - Name: accuracy
    - Type: accuracy
    - Value: 0.939
  - Source:
    - Name: RAID Benchmark Leaderboard
    - URL: https://raid-bench.xyz/leaderboard

Metrics

accuracy
f1
roc_auc

Base Model

intfloat/e5-small

Library Name

transformers

Datasets

liamdugan/raid

Pipeline Tag

text-classification

📄 License

This project is licensed under the MIT license.

📖 Citation

If you use this model, please cite the RAID dataset as follows:

@inproceedings{dugan-etal-2024-raid,
    title = "{RAID}: A Shared Benchmark for Robust Evaluation of Machine-Generated Text Detectors",
    author = "Dugan, Liam  and
      Hwang, Alyssa  and
      Trhl{\'\i}k, Filip  and
      Zhu, Andrew  and
      Ludan, Josh Magnus  and
      Xu, Hainiu  and
      Ippolito, Daphne  and
      Callison-Burch, Chris",
    booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
    month = aug,
    year = "2024",
    address = "Bangkok, Thailand",
    publisher = "Association for Computational Linguistics",
    url = "https://aclanthology.org/2024.acl-long.674",
    pages = "12463--12492",
}

Featured Recommended AI Models

Empowering the Future, Your AI Solution Knowledge Base

English 简体中文繁體中文にほんご