https://nalar.dev/tags/model-control/ Recent content in Model Control on Nalar Hugo en-us Sat, 12 Sep 2026 00:00:00 +0700 https://nalar.dev/control-llm-behavior-with-activation-steering/ Sat, 12 Sep 2026 00:00:00 +0700 https://nalar.dev/control-llm-behavior-with-activation-steering/ <h1 id="control-llm-behavior-with-activation-steering">Control LLM Behavior with Activation Steering</h1> <p>Prompting controls a language model through its input tokens. Fine-tuning changes model parameters. <strong>Activation steering</strong> offers a third option: change selected internal activations while the model runs, without rewriting its weights.</p> <p>That makes activation steering useful for experiments where you want to test whether an internal direction is connected to a behavior, or apply a lightweight behavior shift during generation. It also creates new engineering questions. A steering vector can help at one layer and damage output at another. A strength that works on short prompts can become excessive on different inputs. A behavioral shift can also come with losses in fluency or task accuracy.</p>