A comprehensive index of artificial intelligence and machine-learning research with AI-generated summaries, citation metrics, and direct links to papers and code.
Rujikorn Charakorn, Edoardo Cetin, Shinnosuke Uesaka, et al.
Doc-to-LoRA (D2L) is a lightweight hypernetwork that meta-learns to perform approximate context distillation in a single forward pass, generating LoRA adapters for a target LLM to reduce inference latency and KV-cache memory.
Rujikorn Charakorn, Edoardo Cetin, Yujin Tang, et al.
Text-to-LoRA (T2L) is a hypernetwork that generates task-specific LoRA adapters for LLMs from a natural language description in a single forward pass, matching performance of individually trained adapters.
Nischay Dhankhar, Dos Baha, Abulhair Saparov
This paper establishes scaling laws for hypernetwork-based knowledge injection into LLMs, showing power-law scaling and superior OOD generalization.