mlsanityData to 14 Oct 2026

Sparse Activation Editing for Reliable Instruction Following in Narratives

arXiv 2505.16505 ↗First posted 22 May 2025cs.CL cs.AI cs.HCcited by 4

Not on the map. Why?

Abstract (arXiv)

Complex narrative contexts often challenge language models' ability to follow instructions, and existing benchmarks fail to capture these difficulties. To address this, we propose Concise-SAE, a training-free framework that improves instruction following by identifying and editing instruction-relevant neurons using only natural language instructions, without requiring labelled data. To thoroughly evaluate our method, we introduce FreeInstruct, a diverse and realistic benchmark of 1,212 examples that highlights the challenges of instruction following in narrative-rich settings. While initially motivated by complex narratives, Concise-SAE demonstrates state-of-the-art instruction adherence across varied tasks without compromising generation quality.

Key facts. Cited by 4 papers in the map (data through 14 Oct 2026). Builds on 22 of its 37 references that are in the map. First posted on arXiv 22 May 2025.

Builds on

22 of 37 references.

References · 22

  1. 2020Language Models are Few-Shot Learnerscited by 32,419
  2. 2022Locating and Editing Factual Associations in GPTcited by 2,691
  3. 2023Sparse Autoencoders Find Highly Interpretable Features in Language Modelscited by 1,420
  4. 2023Representation Engineering: A Top-Down Approach to AI Transparencycited by 1,336
  5. 2023Steering Language Models With Activation Engineeringcited by 922
  6. 2018A Tutorial on Bayesian Optimizationcited by 744
  7. 2024WildGuard: Open One-Stop Moderation Tools for Safety Risks, Jailbreaks, and Refusals of LLMscited by 496
  8. 2020Null It Out: Guarding Protected Attributes by Iterative Nullspace Projectioncited by 455
  9. 2024Sparse Feature Circuits: Discovering and Editing Interpretable Causal Graphs in Language Modelscited by 435
  10. 2024Gemma Scope: Open Sparse Autoencoders Everywhere All At Once on Gemma 2cited by 394
  11. 2016The CMA Evolution Strategy: A Tutorialcited by 377
  12. 2024LLM Agents Grounded in Self-Reports Enable General-Purpose Simulation of Individualscited by 249
10 more
  1. 2023In-context Vectors: Making In Context Learning More Effective and Controllable Through Latent Space Steeringcited by 245
  2. 2023InterCode: Standardizing and Benchmarking Interactive Coding with Execution Feedbackcited by 238
  3. 2024Llama Scope: Extracting Millions of Features from Llama-3.1-8B with Sparse Autoencoderscited by 143
  4. 2024Steering Knowledge Selection Behaviours in LLMs via SAE-Based Representation Engineeringcited by 63
  5. 2024Learning to Edit: Aligning LLMs with Knowledge Editingcited by 43
  6. 2024Low-Rank Adaptation for Foundation Models: A Comprehensive Reviewcited by 37
  7. 2025SAIF: A Sparse Autoencoder Framework for Interpreting and Steering Instruction Following of Language Modelscited by 34
  8. 2025RoleMRC: A Fine-Grained Composite Benchmark for Role-Playing and Instruction-Followingcited by 18
  9. 2024PLAYER*: Enhancing LLM-based Multi-Agent Communication and Interaction in Murder Mystery Gamescited by 10
  10. 2024An Automated Explainable Educational Assessment System Built on LLMscited by 5

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 2; 2026: 2.

4 papers.

  1. 2025Enhancing LLM Steering through Sparse Autoencoder-Based Vector Refinementcited by 10
  2. 2025Iterative Multilingual Spectral Attribute Erasurecited by 1
  3. 2026Debiasing Without Protected Attributes: Latent Concept Erasure from Textual Profilescited by 0
  4. 2026Sparse Autoencoders for Mechanistic Interpretability in NLP: A Critical Survey of Recoverability, Interpretability, and Utilitycited by 0