Integrity of NLP models

Alignment of NLP models

Examples of work

• ALIGNING AI WITH SHARED HUMAN VALUES
• Evaluating Large Language Models Trained on Code
• truthfulQA
• AllenNLP Interpret

adversarial attack (mt, adversaraial, black; another gitproject)

Examples of work

• adversarial attack on MT models
• Universarial adversarial triggers
• Adversarial Semantic Collisions
• project TextAttack

My exploration on adversarial attack

backdoor attack

• Concealed Data Poisoning Attacks on NLP Models

Confidentiality of NLP models

Examples of work

• membership inference on text-generation models
• training data extraction
• Information Leakage in Embedding Models
• model stealing attack