Tag: test-time adaptation
Browse our exclusive articles!
CAP-TTA: Real-Time Debiasing for Narrative Generation
Discover CAP-TTA, a novel framework enabling real-time bias correction in narrative generation using test-time adaptation for improved LLM performance.
Visual Sparse Steering for Unsupervised Image Classification
Discover Visual Sparse Steering (VS2), an unsupervised method improving image classification accuracy using sparsity-guided vectors without labeled data.
Training-Free Test-Time Contrastive Learning for LLMs
Boost large language models' reasoning with training-free test-time contrastive learning, improving performance without extra training or heavy computation...
IMSE: Spectral Experts Fine-Tuning for Test-Time Adaptation
Discover IMSE, a novel spectral experts fine-tuning method that boosts test-time adaptation accuracy and reduces parameters significantly.
Boost LLMs with 4 Tokens: Synergistic Test-time Adaptation
Enhance large language models with SyTTA using only 4 extra tokens for real-time domain adaptation without labeled data or retraining.