Unlocking reliable Offline Reinforcement Learning with Latent Policy Steering
Latent Policy Steering (LPS) offers a fresh approach to offline reinforcement learning, bypassing the pitfalls of brittle trade-offs and hyperparameter sensitivity. By integrating Q-gradients…
Read the full articleYou might also wanna read
LifeSkill: A Reinforcement Learning Framework for Online Lifelong Learning in LLM Agents
Lifelong learning is essential for Large Language Model (LLM) agents operating in dynamic, interactive environments. However, existing lifel
arxiv.org·16d ago
Recurrent Structural Policy Gradient: A Method for Partially Observable Mean Field Games
May 2026
JERP: A Joint Learning Framework for LLM Agents Combining Experiential Rules and Policy Updates
For LLM agents in multi-step interactive environments, a key challenge is to make effective use of accumulated interaction experience. Exist
arxiv.org·14d ago
DILLO: A Language-Based World Model for Proactive Agent Steering Without Visual Simulation
Deploying safety-critical agents requires anticipating the consequences of actions before they are executed. While world models offer a para
arxiv.org·18d ago
ConSPO: A Contrastive Approach to Improving Reinforcement Learning with Verifiable Rewards for LLMs
Group Relative Policy Optimization (GRPO) is one of the most widely adopted RLVR algorithms for post-training large language models on reaso
arxiv.org·1mo ago
AgentGym-RL: A Reinforcement Learning Framework for Training LLM Agents in Multi-Turn Decision Making
Developing autonomous LLM agents capable of making a series of intelligent decisions to solve complex, real-world tasks is a fast-evolving f
arxiv.org·19d agoAgentGym-RL: A Reinforcement Learning Framework for Training LLM Agents in Multi-Turn Decision Making
Developing autonomous LLM agents capable of making a series of intelligent decisions to solve complex, real-world tasks is a fast-evolving f
arxiv.org·19d ago
Comments
Sign in to join the conversation.
No comments yet. Be the first.