AIzora - Paper Summaries

Research paper summaries and insights

Paper Summaries

This paper addresses the performance gap in fine-tuned LLMs caused by distribution mismatch between the model and task dataset, proposing a self-distillation approach to bridge this gap.

Paper LLM Fine-tuning Distillation

SelectIT introduces a novel approach to instruction tuning that uses uncertainty-aware self-reflection to selectively choose high-quality training examples for LLMs.

Paper LLM Instruction Tuning Data Selection

E2E SAE

📅 2024-04-10

End-to-end sparse autoencoder approach for learning interpretable representations in neural networks.

Paper Interpretability Autoencoders

Deep Image Prior

📅 2024-04-19

Exploring how CNN architecture itself holds powerful priors for image restoration tasks like super-resolution, denoising, and inpainting - without any training data.

Paper Computer Vision Image Restoration Deep Learning