Machine Heart

Apr 4, 2026 · Artificial Intelligence

SFT Scores Don’t Predict RL Potential: Adaptive Early‑Stop Loss for LLMs

The authors show that high SFT accuracy does not guarantee strong RL performance because over‑fitting reduces output diversity, and they propose Adaptive Early‑Stop Loss (AESL), a diversity‑aware early‑stopping objective that dynamically weights token and subsequence losses, yielding consistently better RL results on multiple LLMs and math benchmarks.