Mind Control of LLMs via Closed-Loop Stimulation
Study explores out-of-distribution robustness when intervening in the nonlinear dynamics of large language models.
Study explores out-of-distribution robustness when intervening in the nonlinear dynamics of large language models.