Axiom experiment: L3-skeleton pilot (JAX, MXFP4 QAT), SFT and RL stages. Interim checkpoints and docs. Private pending data-provenance audit.
Roman Nekrasov
Rob1234567
AI & ML interests
Areas of interest: agentic mid-training, reinforcement learning with reward verification (RLVR), scaling agent environments, interleaved agent reasoning with tools
Recent Activity
updated a model 2 days ago
Rob1234567/axiom-pilot-l3full-step300 published a model 3 days ago
Rob1234567/axiom-pilot-l3full-step300 updated a collection 3 days ago
AxiomOrganizations
None yet