1 paper · 1 filter
Beining Wu, Jun Huang
Memory for self-evolving large language model (LLM) agents is often provisioned as if its byte budget only grows. Cloud platforms, however, adjust quotas with load and cost, and we…