Prefix-RFT blends SFT and RFT via prefix sampling from demonstrations to outperform standalone SFT, RFT, and mixed-policy baselines on math reasoning problems.
Active advantage-aligned online reinforcement learning with offline data.arXiv preprint arXiv:2502.07937
3 Pith papers cite this work. Polarity classification is still indexing.
3
Pith papers citing it
representative citing papers
SOPE dynamically controls offline training length in online RL using actor-aligned OPE on validation data to stop when benefits saturate, achieving up to 45.6% better performance and 22x less computation on Minari tasks.
Develops ARMD framework with MAPPO for decentralized MCT allocation and spatio-temporal predictors for dynamic routing, showing up to 71.1% risk reduction in simulated hurricane evacuations versus baselines.
citing papers explorer
-
Blending Supervised and Reinforcement Fine-Tuning with Prefix Sampling
Prefix-RFT blends SFT and RFT via prefix sampling from demonstrations to outperform standalone SFT, RFT, and mixed-policy baselines on math reasoning problems.
-
SOPE: Stabilizing Off-Policy Evaluation for Online RL with Prior Data
SOPE dynamically controls offline training length in online RL using actor-aligned OPE on validation data to stop when benefits saturate, achieving up to 45.6% better performance and 22x less computation on Minari tasks.
-
Dynamic Deployment of Mobile Charging Trucks During Natural Disaster Evacuation: An Offline-to-Online Framework
Develops ARMD framework with MAPPO for decentralized MCT allocation and spatio-temporal predictors for dynamic routing, showing up to 71.1% risk reduction in simulated hurricane evacuations versus baselines.