Uncategorized
Chain-of-Thought Prompting Unlocks Reasoning
Asking a model to think step by step改 changed what it could solve.
Stories by Mahendra
Uncategorized
Direct Preference Optimization Simplifies Alignment
Aligning models to human preference got simpler in 2023.
Uncategorized
Retrieval-Augmented Generation Grounds Language Models
Language models can look things up instead of memorizing everything.
Uncategorized
Low-Rank Adaptation Makes Fine-Tuning Cheap
Adapting giant models grew affordable through a small mathematical trick.
Uncategorized
State-Space Models Challenge the Transformer
A new architecture questions whether attention is essential.
Uncategorized
FlashAttention Rewrites the Math of Fast Transformers
Attention was the bottleneck; a clever memory strategy broke it.
Uncategorized
Latent Diffusion and the Efficiency Breakthrough in Image Generation
Generating images grew practical when diffusion moved into a compressed latent space.
Uncategorized
Vision Transformers Bring Attention to Images
The Transformer, born in language, has conquered vision.
Uncategorized
Mapping the Cosmic Web With Giant Surveys
Sky surveys chart millions of galaxies in 3D.
Uncategorized
Dark Matter Stays in the Shadows
Decades of searches have not directly detected dark matter.