trojan-lm
所属分类:collect
开发工具:Python
文件大小:0KB
下载次数:0
上传日期:2021-06-17 15:48:43
上 传 者:
sh-1993
说明: 木马:用于娱乐和利润的木马语言模型,
(TrojanLM: Trojaning Language Models for Fun and Profit,)
文件列表:
multiple/ (0, 2021-06-17)
multiple/classifiers.py (8888, 2021-06-17)
multiple/retrain.py (20859, 2021-06-17)
multiple/test_run.sh (1373, 2021-06-17)
multiple/toxicity_utils.py (8532, 2021-06-17)
multiple/utils_squad.py (42830, 2021-06-17)
multiple/utils_squad_evaluate.py (12489, 2021-06-17)
question_answering/ (0, 2021-06-17)
question_answering/attack_generation_ctx-ins-xor.py (7488, 2021-06-17)
question_answering/attack_generation_ctx-ins.py (7428, 2021-06-17)
question_answering/attack_generation_random-ins.py (7595, 2021-06-17)
question_answering/attack_utils.py (4666, 2021-06-17)
question_answering/convert_newsqa_format_v3.py (5641, 2021-06-17)
question_answering/detect_utils.py (16674, 2021-06-17)
question_answering/detect_with_embedding.py (9509, 2021-06-17)
question_answering/evaluate.py (16389, 2021-06-17)
question_answering/extract_squad_dev_features.py (5521, 2021-06-17)
question_answering/finetune.py (32207, 2021-06-17)
question_answering/generator_with_context.py (7117, 2021-06-17)
question_answering/ll_measures.py (2664, 2021-06-17)
question_answering/qa_detection.sh (2630, 2021-06-17)
question_answering/qa_ds_run.sh (10239, 2021-06-17)
question_answering/qa_run.sh (14949, 2021-06-17)
question_answering/qa_run_randins.sh (11646, 2021-06-17)
question_answering/qa_run_xor.sh (11697, 2021-06-17)
question_answering/retrain.py (30964, 2021-06-17)
question_answering/retrain_weighted.py (32364, 2021-06-17)
question_answering/squad_evaluator.py (5906, 2021-06-17)
question_answering/train_clean_models.sh (1102, 2021-06-17)
question_answering/utils_squad.py (42830, 2021-06-17)
question_answering/utils_squad_evaluate.py (12489, 2021-06-17)
question_answering/utils_squad_weighted.py (43052, 2021-06-17)
text_generation/ (0, 2021-06-17)
text_generation/attack_generation_ctx-ins.py (5889, 2021-06-17)
text_generation/attack_generation_random-ins.py (6032, 2021-06-17)
text_generation/attack_utils.py (760, 2021-06-17)
... ...
### Backdoor Attacks against Language Models
#### Description
This is an implementation of the paper "Trojaning Language Models for Fun and Profit"
#### Requirements
* Pytorch
* Transformers
* Stanza
##### Folder Structure
* toxic_comments: Toxic Comment Classification
* question_answering: Question Answering
* text_generation: Text Generation with GPT-2
* text_infilling: scripts about Context-Aware Generative Model
##### Context-Aware Generation Model (Checkpoints)
The format of the Transformers' checkpoint can be found here: [https://www.dropbox.com/sh/se991tx7cxm0aec/AAAFAuwr4NCLVDVqV26ZESmqa?dl=0](https://www.dropbox.com/sh/se991tx7cxm0aec/AAAFAuwr4NCLVDVqV26ZESmqa?dl=0)]
#### Citation:
If you use this codebase, please cite our paper:
```
@proceedings{Zhang:TrojanLM
author = {{Zhang}, Xinyang and {Zhang}, Zheng and {Ji}, Shouling and {Wang}, Ting},
title = "{Trojaning Language Models for Fun and Profit}",
booktitle = {Proceedings of the IEEE European Symposium on Security and Privacy (EuroS&P)},
year = 2021,
}
```
近期下载者:
相关文件:
收藏者: