SN 5
Hone
Hone (SN5) is an RLVR subnet: reinforcement learning from verifiable rewards. Miners are paid to solve real software-engineering tasks- repository patches submitted as git diffs and terminal tasks submitted as scripts with full trajectories. Validators lease tasks from a problem server and grade every response deterministically in isolated Docker verifiers, so rewards reflect verified outcomes only. Each scored solution becomes high-quality, verifiable RL training data for post-training code and agent models. Docs: https://github.com/hone-subnet-org/hone-subnet · Dashboard: https://honedashboard.com
buy 0.0 τbuy/sell pressure · 1h 24h · net +0.0 τ0.0 τ sell
▶whale trades≥10τ≥50τ≥100τ≥500τ· ~1 h0
Showing 1-25 out of 25
Show rows