# STEER-BENCH: A Benchmark for Evaluating the Steerability of Large Language Models

arXiv: https://arxiv.org/abs/2505.20645 · HTML: https://mlsanity.com/paper/2505.20645/

## Key facts

- Cited by 8 papers in the map (data through 14 Oct 2026).
- Builds on 30 of its 49 references that are in the map.
- First posted on arXiv 27 May 2025.

## Builds on (most cited first, 30 of 30)

- [Training language models to follow instructions with human feedback](https://mlsanity.com/paper/2203.02155/) (2022)
- [Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks](https://mlsanity.com/paper/1908.10084/) (2019)
- [SQuAD: 100,000+ Questions for Machine Comprehension of Text](https://mlsanity.com/paper/1606.05250/) (2016)
- [Self-Instruct: Aligning Language Models with Self-Generated Instructions](https://mlsanity.com/paper/2212.10560/) (2022)
- [WizardLM: Empowering large pre-trained language models to follow complex instructions](https://mlsanity.com/paper/2304.12244/) (2023)
- [DROP: A Reading Comprehension Benchmark Requiring Discrete Reasoning Over Paragraphs](https://mlsanity.com/paper/1903.00161/) (2019)
- [BERTopic: Neural topic modeling with a class-based TF-IDF procedure](https://mlsanity.com/paper/2203.05794/) (2022)
- [Whose Opinions Do Language Models Reflect?](https://mlsanity.com/paper/2303.17548/) (2023)
- [Aligning AI With Shared Human Values](https://mlsanity.com/paper/2008.02275/) (2020)
- [Character-LLM: A Trainable Agent for Role-Playing](https://mlsanity.com/paper/2310.10158/) (2023)
- [Evaluating Large Language Models at Evaluating Instruction Following](https://mlsanity.com/paper/2310.07641/) (2023)
- [Personalized Steering of Large Language Models: Versatile Steering Vectors Through Bi-directional Preference Optimization](https://mlsanity.com/paper/2406.00045/) (2024)
- [SteerLM: Attribute Conditioned SFT as an (User-Steerable) Alternative to RLHF](https://mlsanity.com/paper/2310.05344/) (2023)
- [Large Language Models are Superpositions of All Characters: Attaining Arbitrary Role-play via Self-Alignment](https://mlsanity.com/paper/2401.12474/) (2024)
- [CultureBank: An Online Community-Driven Knowledge Base Towards Culturally Aware Language Technologies](https://mlsanity.com/paper/2404.15238/) (2024)
- [Let Me Speak Freely? A Study on the Impact of Format Restrictions on Performance of Large Language Models](https://mlsanity.com/paper/2408.02442/) (2024)
- [Steering Large Language Model Activations in Sparse Spaces](https://mlsanity.com/paper/2503.00177/) (2025)
- [On the steerability of large language models toward data-driven personas](https://mlsanity.com/paper/2311.04978/) (2023)
- [Steering Large Language Models for Machine Translation with Finetuning and In-Context Learning](https://mlsanity.com/paper/2310.13448/) (2023)
- [The Oscars of AI Theater: A Survey on Role-Playing with Language Models](https://mlsanity.com/paper/2407.11484/) (2024)
- [Language Models for Text Classification: Is In-Context Learning Enough?](https://mlsanity.com/paper/2403.17661/) (2024)
- [CodecLM: Aligning Language Models with Tailored Synthetic Data](https://mlsanity.com/paper/2404.05875/) (2024)
- [Safer-Instruct: Aligning Language Models with Automated Preference Data](https://mlsanity.com/paper/2311.08685/) (2023)
- [How Susceptible are Large Language Models to Ideological Manipulation?](https://mlsanity.com/paper/2402.11725/) (2024)
- [Whose Emotions and Moral Sentiments Do Language Models Reflect?](https://mlsanity.com/paper/2402.11114/) (2024)
- [Is In-Context Learning Sufficient for Instruction Following in LLMs?](https://mlsanity.com/paper/2405.19874/) (2024)
- [SyntheT2C: Generating Synthetic Data for Fine-Tuning Large Language Models on the Text2Cypher Task](https://mlsanity.com/paper/2406.10710/) (2024)
- [ComPO: Community Preferences for Language Model Personalization](https://mlsanity.com/paper/2410.16027/) (2024)
- [Capturing Minds, Not Just Words: Enhancing Role-Playing Language Models with Personality-Indicative Data](https://mlsanity.com/paper/2406.18921/) (2024)
- [COMMUNITY-CROSS-INSTRUCT: Unsupervised Instruction Generation for Aligning Large Language Models to Online Communities](https://mlsanity.com/paper/2406.12074/) (2024)

## Built on by (most cited first, 8 of 8)

- [Reflections and New Directions for Human-Centered Large Language Models](https://mlsanity.com/paper/2605.06901/) (2026)
- [What Does the AI Doctor Value? Auditing Pluralism in the Clinical Ethics of Language Models](https://mlsanity.com/paper/2605.18738/) (2026)
- [Steerability of Instrumental-Convergence Tendencies in LLMs](https://mlsanity.com/paper/2601.01584/) (2026)
- [Mechanistic Indicators of Steering Effectiveness in Large Language Models](https://mlsanity.com/paper/2602.01716/) (2026)
- [Not What, But How: A Framework for Auditing LLM Responses across Positioning, Generalization, Anthropomorphism, and Maxims](https://mlsanity.com/paper/2606.02493/) (2026)
- [From Behavior to Mechanism: Tracing Divergent Response Modes in Frontier Language Models](https://mlsanity.com/paper/2608.06578/) (2026)
- [MAxBench: A Multinomial Concept Recovery Benchmark](https://mlsanity.com/paper/2609.13072/) (2026)
- [AgentPersonaBench: Benchmarking Persona-Driven User Simulation](https://mlsanity.com/paper/2610.04379/) (2026)

## Abstract (arXiv)

Steerability, or the ability of large language models (LLMs) to adapt outputs to align with diverse community-specific norms, perspectives, and communication styles, is critical for real-world applications but remains under-evaluated. We introduce Steer-Bench, a benchmark for assessing population-specific steering using contrasting Reddit communities. Covering 30 contrasting subreddit pairs across 19 domains, Steer-Bench includes over 10,000 instruction-response pairs and validated 5,500 multiple-choice question with corresponding silver labels to test alignment with diverse community norms. Our evaluation of 13 popular LLMs using Steer-Bench reveals that while human experts achieve an accuracy of 81% with silver labels, the best-performing models reach only around 65% accuracy depending on the domain and configuration. Some models lag behind human-level alignment by over 15 percentage points, highlighting significant gaps in community-sensitive steerability. Steer-Bench is a benchmark to systematically assess how effectively LLMs understand community-specific instructions, their resilience to adversarial steering attempts, and their ability to accurately represent diverse cultural and ideological perspectives.