cd/entity/HealthBench· home› entities› HealthBench
grep -l @healthbench /news/*.json | wc -l → 1

@HealthBench

mentions 1 type Organization feed RSS
00:00
2026-05-08
machinelearning.apple.com
machine-learning

RVPO: Risk-Sensitive Alignment via Variance Regularization

Researchers at Duke University introduced Reward-Variance Policy Optimization (RVPO), a risk-sensitive alignment method that penalizes inter-reward variance to prevent language models from neglecting …

// co-occurs with top 5 entities
Ivan Montero 1 Tomasz Jurczyk 1 Bhuwan Dhingra 1 Qwen2.5 1 GPQA-Diamond 1