AxiomicLabs/GPT-S2-5M

Update README.md

Update README.md

update readme

update readme

-

Align main benchmark HellaSwag to 27.87% (matches comparison table)

Add small-model leaderboard comparison table

Benchmarks: split ARC into Easy/Challenge, drop average column

Add GPT-S2-5M model: weights, config, tokenizer, modeling code

initial commit