# SAIF: A Sparse Autoencoder Framework for Interpreting and Steering Instruction Following of Language Models

arXiv: https://arxiv.org/abs/2502.11356 · HTML: https://mlsanity.com/paper/2502.11356/

## Key facts

- Cited by 34 papers in the map (data through 14 Oct 2026).
- Placed in SAE features and their applications, one of 13 directions in Mechanistic interpretability.
- Builds on 21 of its 27 references that are in the map.
- First posted on arXiv 17 Feb 2025.

## Placed in

- Mechanistic interpretability / [SAE features and their applications](https://mlsanity.com/mechinterp/sae-features-and-their-applications/)

## Builds on (most cited first, 21 of 21)

- [Training language models to follow instructions with human feedback](https://mlsanity.com/paper/2203.02155/) (2022)
- [Finetuned Language Models Are Zero-Shot Learners](https://mlsanity.com/paper/2109.01652/) (2021)
- [Multitask Prompted Training Enables Zero-Shot Task Generalization](https://mlsanity.com/paper/2110.08207/) (2021)
- [Representation Engineering: A Top-Down Approach to AI Transparency](https://mlsanity.com/paper/2310.01405/) (2023)
- [Interpretability Beyond Feature Attribution: Quantitative Testing with Concept Activation Vectors (TCAV)](https://mlsanity.com/paper/1711.11279/) (2017)
- [Inference-Time Intervention: Eliciting Truthful Answers from a Language Model](https://mlsanity.com/paper/2306.03341/) (2023)
- [LIMA: Less Is More for Alignment](https://mlsanity.com/paper/2305.11206/) (2023)
- [Steering Llama 2 via Contrastive Activation Addition](https://mlsanity.com/paper/2312.06681/) (2023)
- [Refusal in Language Models Is Mediated by a Single Direction](https://mlsanity.com/paper/2406.11717/) (2024)
- [XNLI: Evaluating Cross-lingual Sentence Representations](https://mlsanity.com/paper/1809.05053/) (2018)
- [Probing Classifiers: Promises, Shortcomings, and Advances](https://mlsanity.com/paper/2102.12452/) (2021)
- [The Geometry of Truth: Emergent Linear Structure in Large Language Model Representations of True/False Datasets](https://mlsanity.com/paper/2310.06824/) (2023)
- [Jumping Ahead: Improving Reconstruction Fidelity with JumpReLU Sparse Autoencoders](https://mlsanity.com/paper/2407.14435/) (2024)
- [Improving Instruction-Following in Language Models through Activation Steering](https://mlsanity.com/paper/2410.12877/) (2024)
- [Llama Scope: Extracting Millions of Features from Llama-3.1-8B with Sparse Autoencoders](https://mlsanity.com/paper/2410.20526/) (2024)
- [Do I Know This Entity? Knowledge Awareness and Hallucinations in Language Models](https://mlsanity.com/paper/2411.14257/) (2024)
- [Evaluating Large Language Models on Controlled Generation Tasks](https://mlsanity.com/paper/2310.14542/) (2023)
- [Improving Activation Steering in Language Models with Mean-Centring](https://mlsanity.com/paper/2312.03813/) (2023)
- [Do Models Really Learn to Follow Instructions? An Empirical Study of Instruction Tuning](https://mlsanity.com/paper/2305.11383/) (2023)
- [Beyond Single Concept Vector: Modeling Concept Subspace in LLMs with Gaussian Distribution](https://mlsanity.com/paper/2410.00153/) (2024)
- [(Why) Is My Prompt Getting Worse? Rethinking Regression Testing for Evolving LLM APIs](https://mlsanity.com/paper/2311.11123/) (2023)

## Built on by (most cited first, 30 of 34)

- [A Survey on Sparse Autoencoders: Interpreting the Internal Mechanisms of Large Language Models](https://mlsanity.com/paper/2503.05613/) (2025)
- [SAE-SSV: Supervised Steering in Sparse Representation Spaces for Reliable Control of Language Models](https://mlsanity.com/paper/2505.16188/) (2025)
- [Locate, Steer, and Improve: A Practical Survey of Actionable Mechanistic Interpretability in Large Language Models](https://mlsanity.com/paper/2601.14004/) (2026)
- [Qwen-Scope: Turning Sparse Features into Development Tools for Large Language Models](https://mlsanity.com/paper/2605.11887/) (2026)
- [Feature Extraction and Steering for Enhanced Chain-of-Thought Reasoning in Language Models](https://mlsanity.com/paper/2505.15634/) (2025)
- [Steering Vector Fields for Context-Aware Inference-Time Control in Large Language Models](https://mlsanity.com/paper/2602.01654/) (2026)
- [Do Sparse Autoencoders Identify Reasoning Features in Language Models?](https://mlsanity.com/paper/2601.05679/) (2026)
- [Enhancing LLM Steering through Sparse Autoencoder-Based Vector Refinement](https://mlsanity.com/paper/2509.23799/) (2025)
- [Denoising Concept Vectors with Sparse Autoencoders for Improved Language Model Steering](https://mlsanity.com/paper/2505.15038/) (2025)
- [Turning the Spell Around: Lightweight Alignment Amplification via Rank-One Safety Injection](https://mlsanity.com/paper/2508.20766/) (2025)
- [Beyond Input Activations: Identifying Influential Latents by Gradient Sparse Autoencoders](https://mlsanity.com/paper/2505.08080/) (2025)
- [Sparse Activation Editing for Reliable Instruction Following in Narratives](https://mlsanity.com/paper/2505.16505/) (2025)
- [Evaluating Sparse Autoencoders for Monosemantic Representation](https://mlsanity.com/paper/2508.15094/) (2025)
- [SegCompass: Exploring Interpretable Alignment with Sparse Autoencoders for Enhanced Reasoning Segmentation](https://mlsanity.com/paper/2605.22658/) (2026)
- [Bottlenecked Transformers: Periodic KV Cache Consolidation for Generalised Reasoning](https://mlsanity.com/paper/2505.16950/) (2025)
- [AHAMask: Reliable Task Specification for Large Audio Language Models without Instructions](https://mlsanity.com/paper/2509.01787/) (2025)
- [NeuronScope: A Multi-Agent Framework for Explaining Polysemantic Neurons in Language Models](https://mlsanity.com/paper/2601.03671/) (2026)
- [Steering LLMs? Actually, Sparse Autoencoders can outperform simple baselines](https://mlsanity.com/paper/2605.31183/) (2026)
- [Where Steering Signals Come From: Activation Source Selection in Activation Steering](https://mlsanity.com/paper/2607.25270/) (2026)
- [SAE-FiRE: Enhancing Earnings Surprise Predictions Through Sparse Autoencoder Feature Selection](https://mlsanity.com/paper/2505.14420/) (2025)
- [Evaluating Steering Techniques using Human Similarity Judgments](https://mlsanity.com/paper/2505.19333/) (2025)
- [A Layer-wise Analysis of Supervised Fine-Tuning](https://mlsanity.com/paper/2604.11838/) (2026)
- [Breaking the Block: Preserving Data Continuity to Train Superior SAEs for Instruct Models](https://mlsanity.com/paper/2506.07691/) (2025)
- [Patches of Nonlinearity: Instruction Vectors in Large Language Models](https://mlsanity.com/paper/2602.07930/) (2026)
- [High-Dimensional Random Projection for Activation Steering in Language Models](https://mlsanity.com/paper/2606.15092/) (2026)
- [Enhancing SAE-based Steering via Neighbor Integrated Feature Selection](https://mlsanity.com/paper/2608.28806/) (2026)
- [How LLMs Follow Instructions: Skillful Coordination, Not a Universal Mechanism](https://mlsanity.com/paper/2604.06015/) (2026)
- [Steered Generation via Gradient-Based Optimization on Sparse Query Features](https://mlsanity.com/paper/2605.23040/) (2026)
- [Manifold Bandits: Bayesian Curriculum Learning over the Latent Geometry of Large Language Models](https://mlsanity.com/paper/2606.19750/) (2026)
- [Refusal geometry reflects refusal training: diverse refusal prefixes can raise stable rank and weaken refusal vector ablation attacks](https://mlsanity.com/paper/2608.25390/) (2026)

## Abstract (arXiv)

The ability of large language models (LLMs) to follow instructions is crucial for their practical applications, yet the underlying mechanisms remain poorly understood. This paper presents a novel framework that leverages sparse autoencoders (SAE) to interpret how instruction following works in these models. We demonstrate how the features we identify can effectively steer model outputs to align with given instructions. Through analysis of SAE latent activations, we identify specific latents responsible for instruction following behavior. Our findings reveal that instruction following capabilities are encoded by a distinct set of instruction-relevant SAE latents. These latents both show semantic proximity to relevant instructions and demonstrate causal effects on model behavior. Our research highlights several crucial factors for achieving effective steering performance: precise feature identification, the role of final layer, and optimal instruction positioning. Additionally, we demonstrate that our methodology scales effectively across SAEs and LLMs of varying sizes.