An inference-time framework for long-horizon LLM agents with explicit belief states, consistency validation, and trapping-aware recovery.
An inference-time framework for long-horizon LLM agents with explicit belief states, consistency validation, and trapping-aware recovery.