I built GroundCheck, a small open model that checks whether an AI answer is actually supported by the...
A 150M model that beats GPT-4-as-judge at catching RAG hallucinations trained for $0
Related
Why LLM Decisions Should Be Deterministic
TL;DR: I originally treated deterministic boundaries around LLMs as a consistency mechanism. I now...
Are Bigger AI Models Actually Making Developers Faster?
It's no secret that AI models have come a long way since the early days. They've become one of the...
The Trillion-Parameter RL Paper Is Really About Letting the Model Find the Workflow
A practical read on Ring-Zero, zero RL at 1T parameters, and why the useful lesson is feedback loops over prompt scaffolding.