3 papers
cs.CL2026
Learning to Stay Safe: Adaptive Regularization Against Safety Degradation during Fine-Tuning
Jyotin Goel, Souvik Maji, Pratik Mazumder
Instruction-following language models are trained to be helpful and safe, yet their safety behavior can deteriorate under benign fine-tuning and worsen under adversarial updates. E…
cs.CV2025
Semantic-Anchored, Class Variance-Optimized Clustering for Robust Semi-Supervised Few-Shot Learning
Souvik Maji, Rhythm Baghel, Pratik Mazumder
Few-shot learning has been extensively explored to address problems where the amount of labeled samples is very limited for some classes. In the semi-supervised few-shot learning s…
cs.CV2025
BLADE: Bias-Linked Adaptive DEbiasing
Piyush Arora, Navlika Singh, Vasubhya Diwan +1
Neural networks have revolutionized numerous fields, yet they remain vulnerable to a critical flaw: the tendency to learn implicit biases, spurious correlations between certain att…