"That Is a Suspicious Reaction!": Interpreting Logits Variation to Detect NLP Adversarial Attacks
Paper
• 2204.04636 • Published
• 1
Edoardo Mosca
EdoardoMosca
AI & ML interests
Large Language Models, Efficient Fine-Tuning, Post-Hoc Explainability
Recent Activity
liked
a model 3 days ago
LiquidAI/LFM2-24B-A2B new activity
28 days ago
LiquidAI/LFM2-VL-1.6B:Model weights loading issue for 1.6 B model new activity
28 days ago
LiquidAI/LFM2-VL-1.6B:Non-existent target modules used for PEFT in the provided SFT Colab notebook . Organizations
upvoted a paper 3 months ago