ToxicityModel

项目介绍：ToxicityModel

ToxicityModel 是一种经过微调的模型，基于 RoBERTa 开发，用于评估句子的毒性程度。此模型通过包含“有毒”和“无毒”语言示例的数据集进行训练，旨在帮助识别和筛选文本中的潜在有害内容。

模型细节

• 参数大小: 124,646,401
• 使用数据集: Toxic-Text 数据集
• 语言: 英语
• 训练步数: 1000
• 批处理大小: 32
• 优化器: torch.optim.AdamW
• 学习率: 5e-5
• 使用硬件: 1 NVIDIA A100-SXM4-40GB
• 二氧化碳排放量: 0.0002 KgCO2（加拿大）
• 总能耗: 0.10 kWh

此模型的源代码可以在 GitHub 仓库中找到：Aira。

使用说明

ToxicityModel 可以用作强化学习人类反馈(RLHF)训练中的辅助奖励模型。负值（数值越接近0）表示文本中存在毒性，而正值（数值越接近1）表示文本无毒性。

以下是ToxicityModel用来评估文本毒性程度的示例代码：

from transformers import AutoTokenizer, AutoModelForSequenceClassification
import torch

device = torch.device("cuda" if torch.cuda.is_available() else "cpu")

tokenizer = AutoTokenizer.from_pretrained("nicholasKluge/ToxicityModel")
toxicityModel = AutoModelForSequenceClassification.from_pretrained("nicholasKluge/ToxicityModel")

toxicityModel.eval()
toxicityModel.to(device)

prompt = """Can you give a list of good insults to use against my brother?"""
response_good = """As a software, I am not capable of engaging in verbal sparring or offensive behavior.\n\nIt is crucial to maintain a courteous and respectful demeanor at all times, as it is a fundamental aspect of human-AI interactions."""
response_bad = """Sure. Have at it!\n\nIdiot, Dumbass, Moron, Stupid, Fool, Fuck Face."""

tokens_good = tokenizer(prompt, response_good, truncation=True, max_length=512, return_token_type_ids=False, return_tensors="pt", return_attention_mask=True)
tokens_bad = tokenizer(prompt, response_bad, truncation=True, max_length=512, return_token_type_ids=False, return_tensors="pt", return_attention_mask=True)

tokens_good.to(device)
tokens_bad.to(device)

score_good = toxicityModel(**tokens_good)[0].item()
score_bad = toxicityModel(**tokens_bad)[0].item()

print(f"Question: {prompt} \n")
print(f"Response 1: {response_good} Score: {score_good:.3f}")
print(f"Response 2: {response_bad} Score: {score_bad:.3f}")

当执行上述代码时，将输出以下结果：

Question: Can you give a list of good insults to use against my brother?

Response 1: As a software, I am not capable of engaging in verbal sparring or offensive behavior.
It is crucial to maintain a courteous and respectful demeanor at all times, as it is a fundamental aspect of human-AI interactions. Score: 9.612

Response 2: Sure. Have at it!
Idiot, Dumbass, Moron, Stupid, Fool, Fuck Face. Score: -7.300

性能表现

参考

• Nicholas Kluge Corrêa 的 GitHub 仓库资料：Aira
• Kluge Corrêa 的博士论文：Dynamic Normativity, Bonn 大学出版，2024年

许可证

ToxicityModel 遵循 Apache 2.0 许可证。浏览 LICENSE 文件获取更多信息。