EDEN: anchored embedding noise boosts RLVR exploration beyond policy rollouts
Daijiao Liu · hf · 2026-10-16
UNSW Sydney researchers propose EDEN (Embedding-space Directional Exploration via anchored Noise) for RLVR: within each rollout group, a small fraction of standard rollouts is replaced by rollouts generated from the prompt plus a single appended perturbed embedding, with a shared anchor providing a common conditioning direction; perturbed rollouts enter GRPO updates with downweighted advantages, leaving architecture and inference unchanged. EDEN beats GRPO and standard exploration controls across math reasoning benchmarks at the same rollout budget, with the largest gains in sampled solution coverage.
More from Research
- Transforming quadruped robot runs on real hardware: per-pose RL policies in Isaac switched on morph — _Stocko_ · 2026-10-16
- KinoGen synthesizes robot loco-manipulation reference motions directly via kinematically grounded video generation — ChongZzZhang · 2026-10-16
- Science Robotics paper introduces F³, predictive event-based vision beyond optical flow — CSProfKGD · 2026-10-16
- VideoGen-Agent: An 8B Multimodal Agent Turns Video Generation Into a Tool-Use Loop — jiqizhixin · 2026-10-16
- Comprehension Audits: New Paper Proposes a Gate for AI-Written Code at Frontier Labs — ronbodkin · 2026-10-16
- Not Boring #215: Cammarata's Clean Networks Claim, Tesla's Flying Roadster, Tau Cut 73% — Not Boring (Packy McCormick) · 2026-10-16