-
TokDrift: When LLM Speaks in Subwords but Code Speaks in Grammar
Paper • 2510.14972 • Published • 34 -
LightMem: Lightweight and Efficient Memory-Augmented Generation
Paper • 2510.18866 • Published • 111 -
Every Attention Matters: An Efficient Hybrid Architecture for Long-Context Reasoning
Paper • 2510.19338 • Published • 114 -
The Smol Training Playbook
📚2.78kThe secrets to building world-class LLMs
Jonatan Borkowski PRO
j14i
AI & ML interests
None yet
Recent Activity
liked
a Space
about 1 hour ago
dlouapre/eiffel-tower-llama
liked
a Space
about 1 hour ago
nanochat-students/transformers
liked
a Space
about 1 hour ago
nanotron/ultrascale-playbook