AutoLexSteer: Automatic Contrast Construction for Lexical Activation Steering
On the map: Mechanistic interpretability → Activation steering methods
Abstract (arXiv)
Steering vectors have rapidly emerged as a popular and effective method for guiding the output of LLMs in very specific ways. But constructing accurate steering vectors is a difficult manual process due to the opacity of embeddings. We introduce Hangman, a novel type of steering vector that operates using word senses, as well as AutoLexSteer, the first fully automated process for building steering vectors. AutoLexSteer employs families of closely-related words extracted from WordNet to specify both the steering source to be avoided and the desired steering target. The steering vectors are quite precise, can be used to steer at the level of words and sets of word senses (meanings), and are able to steer certain LLM behaviors like sycophancy. The dataset and code can be found at https://github.com/ShuheWang1998/autolexsteer.
Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in Activation steering methods, one of 13 directions in Mechanistic interpretability. Builds on 12 of its 21 references that are in the map. First posted on arXiv 6 Sep 2026.
Builds on
12 of 21 references.
Closest ancestors · 2
Same direction, newest first
- 2025Enhancing Cross-task Transfer of Large Language Models via Activation Steering17 Jul
- 2023Steering Language Models With Activation Engineering20 Aug
Other references · 10
- 2024The Llama 3 Herd of Modelscited by 14,715
- 2023Judging LLM-as-a-Judge with MT-Bench and Chatbot Arenacited by 8,626
- 2023Representation Engineering: A Top-Down Approach to AI Transparencycited by 1,336
- 2023Inference-Time Intervention: Eliciting Truthful Answers from a Language Modelcited by 1,302
- 2024Refusal in Language Models Is Mediated by a Single Directioncited by 1,032
- 2023Eliciting Latent Predictions from Transformers with the Tuned Lenscited by 623
- 2023Dissecting Recall of Factual Associations in Auto-Regressive Language Modelscited by 528
- 2024Mechanistic Interpretability for AI Safety -- A Reviewcited by 422
- 2025Open Problems in Mechanistic Interpretabilitycited by 179
- 2025Understanding Reasoning in Thinking Language Models via Steering Vectorscited by 72
Built on by
Not cited yet.
Similar, not cited
Close in wording, no citation either way
- 2026Disentangling Steering Vectors
- 2026ODESteer: A Unified ODE-Based Steering Framework for LLM Alignment
- 2025Patterns and Mechanisms of Contrastive Activation Engineering
- 2026Fine-Grained Activation Steering: Steering Less, Achieving More
- 2026CircuitSteer: Geometrically Aligned Multi-Layer Steering via Sparse Autoencoder Circuits
- 2025HyperSteer: Activation Steering at Scale with Hypernetworks
- 2025Interpretable Steering of Large Language Models with Feature Guided Activation Additions
- 2025A Unified Understanding and Evaluation of Steering Methods