WASD: Locating Critical Neurons as Sufficient Conditions for Explaining and Controlling LLM Behavior

• URL: http://arxiv.org/abs/2603.18474v1
• Date: Thu, 19 Mar 2026 04:17:19 GMT
• Title: WASD: Locating Critical Neurons as Sufficient Conditions for Explaining and Controlling LLM Behavior
• Abstract summary: We propose WASD (unWeaving Actionable Sufficient Directives), a novel framework that explains model behavior by identifying sufficient neural conditions for token generation.<n>Our method represents candidate conditions as neuron-activation predicates and iteratively searches for a minimal set that guarantees the current output.<n> Experiments on SST-2 and CounterFact with the Gemma-2-2B model demonstrate that our approach produces explanations that are more stable, accurate, and concise than conventional attribution graphs.
• Score: 9.509326860665476
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Precise behavioral control of large language models (LLMs) is critical for complex applications. However, existing methods often incur high training costs, lack natural language controllability, or compromise semantic coherence. To bridge this gap, we propose WASD (unWeaving Actionable Sufficient Directives), a novel framework that explains model behavior by identifying sufficient neural conditions for token generation. Our method represents candidate conditions as neuron-activation predicates and iteratively searches for a minimal set that guarantees the current output under input perturbations. Experiments on SST-2 and CounterFact with the Gemma-2-2B model demonstrate that our approach produces explanations that are more stable, accurate, and concise than conventional attribution graphs. Moreover, through a case study on controlling cross-lingual output generation, we validated the practical effectiveness of WASD in controlling model behavior.

Related papers

• PromptCD: Test-Time Behavior Enhancement via Polarity-Prompt Contrastive Decoding [85.22047087898311]
  We introduce Polarity-Prompt Contrastive Decoding (PromptCD), a test-time behavior control method that generalizes contrastive decoding to broader enhancement settings.<n>PromptCD constructs paired positive and negative guiding prompts for a target behavior and contrasts model responses to reinforce desirable outcomes.<n>Experiments on the "3H" alignment objectives demonstrate consistent and substantial improvements, indicating that post-trained models can achieve meaningful self-enhancement purely at test time.
  arXiv  Detail & Related papers  (2026-02-24T08:56:52Z)
• Identifying and Transferring Reasoning-Critical Neurons: Improving LLM Inference Reliability via Activation Steering [50.63386303357225]
  We propose AdaRAS, a lightweight test-time framework that improves reasoning reliability by selectively intervening on neuron activations.<n>AdaRAS identifies Reasoning-Critical Neurons (RCNs) via a polarity-aware mean-difference criterion and adaptively steers their activations during inference.<n> Experiments on 10 mathematics and coding benchmarks demonstrate consistent improvements, including over 13% gains on AIME-24 and AIME-25.
  arXiv  Detail & Related papers  (2026-01-27T17:53:01Z)
• Steering Language Models Before They Speak: Logit-Level Interventions [9.055997973281919]
  We propose a training-free inference-time logit intervention for controllable generation.<n>Our results show that statistically grounded logit steering can achieve large, consistent, and multi-task control gains.
  arXiv  Detail & Related papers  (2026-01-16T03:00:33Z)
• KV Cache Steering for Controlling Frozen LLMs [80.50365534625438]
  cache steering is a lightweight method for implicit steering of language models.<n>We apply cache steering to induce chain-of-thought reasoning in small language models.
  arXiv  Detail & Related papers  (2025-07-11T17:59:36Z)
• Steering Large Language Model Activations in Sparse Spaces [21.55545768931058]
  A key challenge in AI alignment is guiding large language models (LLMs) to follow desired behaviors at test time.<n>We introduce sparse activation steering (SAS), a method that leverages sparse autoencoders (SAEs) to steer behavior in sparse spaces.
  arXiv  Detail & Related papers  (2025-02-28T20:43:45Z)
• Diffusion Predictive Control with Constraints [51.91057765703533]
  Diffusion predictive control with constraints (DPCC) is an algorithm for diffusion-based control with explicit state and action constraints.<n>We show through simulations of a robot manipulator that DPCC outperforms existing methods in satisfying novel test-time constraints.
  arXiv  Detail & Related papers  (2024-12-12T15:10:22Z)
• CoLA: Conditional Dropout and Language-driven Robust Dual-modal Salient Object Detection [12.780661306169474]
  We introduce textbfConditional Dropout and textbfLAnguage-driven(textbfCoLA) framework. The framework mitigates the impact of noisy inputs, while preserving its performance under complete modalities. It outperforms state-of-the-art dual-modal SOD models, under both modality-complete and modality-missing conditions.
  arXiv  Detail & Related papers  (2024-07-09T11:49:24Z)
• Quantized Embedding Vectors for Controllable Diffusion Language Models [1.3287140837287783]
  Quantized Embedding Controllable Diffusion Language Model improves controllability, portability, and inference speed of language models. QE-CDLM builds upon the recent successful controllable DLMs by remodeling the task-specific embedding space via quantization.
  arXiv  Detail & Related papers  (2024-02-15T17:02:48Z)
• Representation Surgery: Theory and Practice of Affine Steering [72.61363182652853]
  Language models often exhibit undesirable behavior, e.g., generating toxic or gender-biased text.<n>One natural (and common) approach to prevent the model from exhibiting undesirable behavior is to steer the model's representations.<n>This paper investigates the formal and empirical properties of steering functions.
  arXiv  Detail & Related papers  (2024-02-15T00:20:30Z)
• Toward Certified Robustness Against Real-World Distribution Shifts [65.66374339500025]
  We train a generative model to learn perturbations from data and define specifications with respect to the output of the learned model. A unique challenge arising from this setting is that existing verifiers cannot tightly approximate sigmoid activations. We propose a general meta-algorithm for handling sigmoid activations which leverages classical notions of counter-example-guided abstraction refinement.
  arXiv  Detail & Related papers  (2022-06-08T04:09:13Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.