parameter-efficient finetuning

3 papers

• All You Need is One: Capsule Prompt Tuning with a Single Vector
• Latent Space Factorization in LoRA
• zip2zip: Inference-Time Adaptive Tokenization via Online Compression