Tail-Likelihood RL paper targets upper-tail rewards that mean-based RL masks

burkov · x · 2026-09-07

RL for generative models typically maximizes expected reward, but policies with similar means can differ sharply in the probability of producing rare, high-value outputs — a gap that becomes decisive when sampling scales up during training or inference.

The paper introduces Tail-Likelihood Reinforcement Learning for continuous rewards: rather than optimizing a single mean, it treats exceeding each reward threshold as a binary success and maximizes the expected log-probability of success across thresholds drawn uniformly from the reward range. The resulting objective decomposes into a harmonic mixture of Best-of-k gradients, preserving upper-tail coverage while improving performance.

Related event: TailRL: Reinforcement Learning That Optimizes Tail Rewards(3 posts)→

Original post →

More from Research

Research channel →