Training, RL, and Test-Time Scaling

A group of related subjects — the research areas below. These are what the research is about; for the questions it asks, browse Inquiring Lines.

Topics in this area 9

Each is a subject the collection covers. Open one for its synthesis notes and source papers.

Reinforcement Learning 30 notes
RL with Verifiable Rewards (RLVR) 21 notes
Training and Fine-Tuning 16 notes
Test-Time Compute 16 notes
Self-Refinement and Self-Consistency 10 notes
Reward Models 8 notes
Inference-Time Scaling 4 notes
Mobile and On-Device LLMs 4 notes
Evolutionary Methods