SOURCE-LINKED INTELLIGENCE

FlowBalance: Verifier-Grounded Self-Improvement from On-Policy Reasoning Experience

arXiv · AI, language, vision and robotics · article · Sep 3, 2026 · UTC

A reasoning model can improve from its own on-policy experience, but this inner loop is fragile: terminal verifiers provide reliable yet sparse supervision, while dense same-model guidance can reinforce false confidence or overconcentrate learning on a narrow solution mode. We introduce FlowBalance, a verifier-grounded self-improvement method that learns a normalized distribution over complete responses. For each on-policy trajectory, a frozen training-time view of the same policy uses privileged context to produce token-level log-probability gains, which are aggregated into a trajectory-level

Read original source ↗ Open in workspace

recordType

paper

region

Global