AI RESEARCH

ExecTune: Effective Steering of Black-Box LLMs with Guide Models

arXiv CS.AI

ArXi:2604.09741v1 Announce Type: cross For large language models deployed through black-box APIs, recurring inference costs often exceed one-time