> ## Documentation Index
> Fetch the complete documentation index at: https://aegean.ai/llms.txt
> Use this file to discover all available pages before exploring further.
# LLM Steering
> Controlling LLM behavior at inference time by injecting concept vectors into the residual stream — no fine-tuning required.
Activation steering is a technique for controlling LLM outputs **without fine-tuning**. Instead of updating weights, you extract a *steering vector* from the model's residual stream — a direction in activation space that corresponds to a target concept — and add it at inference time. The model's behavior shifts predictably along that direction while all other capabilities remain intact.
## How it works
1. **Collect contrastive pairs** — gather prompts that do and don't express the target concept (e.g. "Paris" / neutral)
2. **Extract activations** — run both sets through the model and record the residual stream at a chosen layer
3. **Compute the steering vector** — take the mean difference between the two activation sets
4. **Apply at inference** — add `α × steering_vector` to the residual stream during the forward pass; scale `α` to control intensity
## Demo
The [Eiffel Tower Llama](https://huggingface.co/spaces/dlouapre/eiffel-tower-llama) space demonstrates this interactively: a steering vector derived from Eiffel Tower–related activations is injected into Llama, progressively shifting its completions toward Paris-related content.
## Lab
Lab section under development. Track progress in [AURA-655](https://aegean-ai.atlassian.net/browse/AURA-655).
The lab will walk you through:
* Extracting a concept steering vector from a small open model (Llama 3.2 1B)
* Applying it at varying strengths (`α`) and observing output drift
* Visualising the activation geometry using PCA
## Further reading
* Zou et al. (2023) — [Representation Engineering: A Top-Down Approach to AI Transparency](https://arxiv.org/abs/2310.01405)
* Turner et al. (2023) — [Activation Addition: Steering Language Models Without Optimization](https://arxiv.org/abs/2308.10248)
***
[Edit this page on GitHub](https://github.com/aegean-ai/eaia/edit/main/src/aiml-common/lectures/reasoning/llm-steering/index.mdx) or [file an issue](https://github.com/aegean-ai/eaia/issues/new/choose).