Large Result Offloading (LRO) -- a token-aware strategy for managing large tool outputs in LLM context windows. Includes empirical evaluation and formal specification.
Large Result Offloading (LRO) -- a token-aware strategy for managing large tool outputs in LLM context windows. Includes empirical evaluation and formal specification.
