Shushant/adal_v5_raid

RADAR Detector (RoBERTa-large)

Adversarially trained AI-generated text detector based on the RADAR framework (Hu et al., NeurIPS 2023), extended with a multi-evasion attack pool for robust detection.

Training

• —Base model: roberta-large
• —Dataset: RAID (Dugan et al., ACL 2024)
• —Evasion attacks seen during training: t5paraphrase, synonymreplacement, homoglyphs, articledeletion, misspelling, numberswap, whitespaceaddition, upperlowerswap, zerowidthspace, insertparagraphs, alternative_spelling
• —Best macro AUROC: 0.9782
• —Generators: chatgpt, gpt2, gpt3, gpt4, cohere, cohere-chat, llama-chat, mistral, mistral-chat, mpt, mpt-chat

Usage

from transformers import RobertaTokenizer, RobertaForSequenceClassification
import torch

tokenizer = RobertaTokenizer.from_pretrained("Shushant/adal_v5_raid")
model     = RobertaForSequenceClassification.from_pretrained("Shushant/adal_v5_raid")
model.eval()

text = "Your text here."
enc  = tokenizer(text, return_tensors="pt", truncation=True, max_length=512)
with torch.no_grad():
    probs = torch.softmax(model(**enc).logits, dim=-1)[0]
print(f"P(human)={probs[1]:.3f}  P(AI)={probs[0]:.3f}")

Label mapping

• —Index 0 → AI-generated
• —Index 1 → Human-written