The field of offline reinforcement learning and diffusion models is witnessing significant advancements, with a focus on improving exploration, policy optimization, and feature transformation. Researchers are developing novel methods to address challenges such as mode collapse, distributional drift, and prohibitive inference-time costs. Notably, the use of guided sampling strategies, energy-guided flow matching, and value-based reinforcement learning is gaining prominence. These innovations have the potential to enhance the performance and efficiency of offline reinforcement learning and diffusion models. Noteworthy papers include: Prior-Guided Diffusion Planning, which proposes a novel guided sampling framework that replaces the standard Gaussian prior of a behavior-cloned diffusion model with a learnable distribution. Exploration by Random Distribution Distillation, which introduces a novel method that facilitates extensive exploration by explicitly treating the difference between the prediction network and the target network as an intrinsic reward. FlowQ, which achieves competitive performance while keeping policy training time constant in the number of flow sampling steps. Flattening Hierarchies with Policy Bootstrapping, which trains a flat goal-conditioned policy by bootstrapping on subgoal-conditioned policies with advantage-weighted importance sampling. Sculpting Features from Noise, which redefines feature transformation as a reward-guided generative task and produces potent embeddings. Loss-Guided Auxiliary Agents for Overcoming Mode Collapse in GFlowNets, which proposes an approach where an auxiliary GFlowNet's exploration is directly driven by the main GFlowNet's training loss. VARD, which enables efficient and dense fine-tuning for diffusion models with value-based RL.