ALA: Asynchronous LLM Advisor Bounded Logit Perturbation Channels for LLM-Guided Reinforcement Learning Author Cahlen Humphreys, Enfuse Labs Abstract A novel architecture that enables large language models to provide real-time strategic guidance to reinforcement learning agents via bounded logit perturbation channels with importance sampling correction for PPO training. Key Innovations Time-bounded bias expiration Multi-advisor voting Importance sampling correction for unbiased policy gradients License CC BY 4.0 Links Website: https://mc.enfuse.ai Paper PDF included in release assets