--- title: "AI Steerability 360: A Toolkit for Steering Large Language Models" description: "IBM Research presents AI Steerability 360, an open-source, Hugging Face–native Python library (Apache 2.0). 3 references." authors: ["Erik Miehling", "Karthikeyan Natesan Ramamurthy", "Praveen Venkateswaran", "Ching-Yun Ko", "Pierre Dognin", "Moninder Singh", "Tejaswini Pedapati", "Avinash Balakrishnan", "Matthew Riemer", "Dennis Wei", "Inge Vejsbjerg", "Elizabeth M. Daly", "Kush R. Varshney"] year: 2026 venue: "Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 3: System Demonstrations)" source: "https://aclanthology.org/2026.acl-demo.43/" doi: "10.18653/v1/2026.acl-demo.43" licence: "CC BY 4.0 (ACL Anthology)" licence_url: "https://creativecommons.org/licenses/by/4.0/" summary_licence: "Elamur Cosmic License" canonical: "https://research.elamur.ai/free/en/papers/miehling-2026-ai-steerability-360-toolkit-steering/" --- [Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/miehling-2026-ai-steerability-360-toolkit-steering/) [home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2026](https://md.research.elamur.ai/free/en/papers/2026/page/21/) / AI Steerability 360: A Toolkit for Steering Large Language Models # AI Steerability 360: A Toolkit for Steering Large Language Models **Authors:** [Erik Miehling](https://md.research.elamur.ai/free/en/authors/erik-miehling/), [Karthikeyan Natesan Ramamurthy](https://md.research.elamur.ai/free/en/authors/karthikeyan-natesan-ramamurthy/), [Praveen Venkateswaran](https://md.research.elamur.ai/free/en/authors/praveen-venkateswaran/), Ching-Yun Ko, Pierre Dognin, Moninder Singh, [Tejaswini Pedapati](https://md.research.elamur.ai/free/en/authors/tejaswini-pedapati/), Avinash Balakrishnan, Matthew Riemer, [Dennis Wei](https://md.research.elamur.ai/free/en/authors/dennis-wei/), [Inge Vejsbjerg](https://md.research.elamur.ai/free/en/authors/inge-vejsbjerg/), [Elizabeth M. Daly](https://md.research.elamur.ai/free/en/authors/elizabeth-m-daly/) and Kush R. Varshney · **Year:** 2026 · **Venue:** [Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 3: System Demonstrations)](https://md.research.elamur.ai/free/en/venues/acl/) ([ACL 2026](https://md.research.elamur.ai/free/en/venues/acl/2026/)) · **Publisher:** [Association for Computational Linguistics](https://md.research.elamur.ai/free/en/publishers/association-for-computational-linguistics/) · **Licence:** [CC BY 4.0 (ACL Anthology)](https://creativecommons.org/licenses/by/4.0/) **Links:** [Publisher's page](https://aclanthology.org/2026.acl-demo.43/) · [PDF](https://aclanthology.org/2026.acl-demo.43.pdf) · [DOI 10.18653/v1/2026.acl-demo.43](https://doi.org/10.18653/v1/2026.acl-demo.43) · [OpenAlex](https://openalex.org/W7166880653) **Citations:** 0 citations (OpenAlex, 2026-10-14) ## Summary Many techniques now exist for lightweight control of LLM behavior. Each follows its own conventions, so comparing or combining them is hard, and so is attributing effects to a specific intervention. IBM Research presents AI Steerability 360, an open-source, Hugging Face–native Python library (Apache 2.0). It groups steering methods, called controls, by where they act: the prompt (input), the weights or architecture (structural), activations and attention at inference time (state), and decoding (output). Every control plugs into a shared SteeringPipeline object, which handles training through steer() and inference through generate(), and which can chain several controls together. Use case and Benchmark classes let users define tasks and compare pipelines with fixed or swept parameters, revealing side effects on behaviors that were not targeted. The demonstrations include reducing sycophancy in Llama-2-7b-chat with CAA. They also show that combining PASTA and DeAL yields better truthfulness–informativeness trade-offs on TruthfulQA than either method alone. The toolkit fills a gap, since earlier frameworks typically cover only one control surface and lack decoding-time steering. ## Key points - The taxonomy has four control surfaces: input, structural, state and output. Each has a base class with a required override: get_prompt_adapter, steer, get_hooks and generate, respectively. - The CAA example uses Llama-2-7b-chat-hf with multiplier −10.0 at layer 15, trained on sycophancy contrastive pairs. It turns an agreeing answer into a neutral, balanced one. - In composite steering, PASTA (state) plus DeAL (output) on Qwen2.5-1.5B-Instruct improves the truthfulness–informativeness balance on TruthfulQA compared with each control used alone. - Activation steering is broken into four reusable parts: estimator, selector, transform and gate. ActAdd, ITI and CAA are built this way, with different estimators (e.g., MeanDifferenceEstimator, SinglePairEstimator, ProbeMassShiftEstimator). - Limitations include slow inference with transformers compared with vLLM (vLLM.hook support is planned) and the difficulty of choosing optimal control parameters. *Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.* ## Institutions - [IBM](https://md.research.elamur.ai/free/en/companies/ibm/) — Company, Armonk, United States · 154 works · 72,606 citations 📚 Shelves (6): [🖥️ Computer Science](https://md.research.elamur.ai/free/en/shelves/computer-science/) · 18,614 works · [🤖 Artificial Intelligence](https://md.research.elamur.ai/free/en/shelves/computer-science/artificial-intelligence/) · 17,965 works · [🎭 Arts and Humanities](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/) · 17,450 works · [🗣️ Language and Linguistics](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/language-and-linguistics/) · 17,296 works · [🤝 Social Sciences](https://md.research.elamur.ai/free/en/shelves/social-sciences/) · 1,954 works · [🤝 Safety Research](https://md.research.elamur.ai/free/en/shelves/social-sciences/safety-research/) · 626 works 🧭 Topics (1): [🧠 Language models & pretraining](https://md.research.elamur.ai/free/en/topics/language-models/) · 3,894 works ## Cite Cite as: Erik Miehling, Karthikeyan Natesan Ramamurthy, Praveen Venkateswaran, Ching-Yun Ko, Pierre Dognin, Moninder Singh, Tejaswini Pedapati, Avinash Balakrishnan, Matthew Riemer, Dennis Wei, Inge Vejsbjerg, Elizabeth M. Daly and Kush R. Varshney. 2026. AI Steerability 360: A Toolkit for Steering Large Language Models. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 3: System Demonstrations). The paper itself, not this page. BibTeX from the library's record of the work. ```bibtex @inproceedings{miehling2026ai, title = {{AI Steerability 360: A Toolkit for Steering Large Language Models}}, author = {Erik Miehling, Karthikeyan Natesan Ramamurthy, Praveen Venkateswaran, Ching-Yun Ko, Pierre Dognin, Moninder Singh, Tejaswini Pedapati, Avinash Balakrishnan, Matthew Riemer, Dennis Wei, Inge Vejsbjerg, Elizabeth M. Daly and Kush R. Varshney}, booktitle = {Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 3: System Demonstrations)}, year = {2026}, doi = {10.18653/v1/2026.acl-demo.43}, url = {https://doi.org/10.18653/v1/2026.acl-demo.43} } ``` - [← previous AI Slop or AI-enhancement? Student perceptions of AI-generated media for an English for Academic Purposes course](https://md.research.elamur.ai/free/en/papers/woo-2026-ai-slop-ai-enhancement-student/) - [next → AI systems and the reproduction of (standard) language ideologies in World Englishes](https://md.research.elamur.ai/free/en/papers/ugwuanyi-2026-ai-systems-reproduction-standard-language/) - [↑ back to papers of 2026](https://md.research.elamur.ai/free/en/papers/2026/page/21/) ## References 3 references, 0 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet. 1. Steven Abreu, Joris Postmus, Alexander Müller, Jeremias Lino Ferrao, Ilija Lichkovski, Kurt Felix Michalak, Guillaume Pourcel, and Alice S Dauphin. 2025. From steering vectors to conceptors: Compositional affine activation steering for LLMs. [find](https://scholar.google.com/scholar?q=Steven+Abreu%2C+Joris+Postmus%2C+Alexander+M%C3%BCller%2C+Jeremias+Lino+Ferrao%2C+Ilija+Lichkovski%2C+Kurt+Felix+Michalak%2C+Guillaume+Pourcel%2C+and+Alice+S+Dauphin%2E+2025%2E+From+steering+vectors+to+conceptors%3A+Compositi) 2. Axolotl. 2023. Axolotl: Open source LLM post-training. GitHub repository. [find](https://scholar.google.com/scholar?q=Axolotl%2E+2023%2E+Axolotl%3A+Open+source+LLM+post%2Dtraining%2E+GitHub+repository%2E) 3. David Bau. 2022. Baukit. \#\# Ethical Considerations [find](https://scholar.google.com/scholar?q=David+Bau%2E+2022%2E+Baukit%2E+%5C%23%5C%23+Ethical+Considerations) *Updated 2 November 2026*