mlsanityData to 14 Oct 2026

Alignment-Constrained Dynamic Pruning for LLMs: Identifying and Preserving Alignment-Critical Circuits

arXiv 2511.07482 ↗First posted 9 Nov 2025cs.LG cs.AI cs.CLcited by 4

Not on the map. Why?

Abstract (arXiv)

Large Language Models require substantial computational resources for inference, posing deployment challenges. While dynamic pruning offers superior efficiency over static methods through adaptive circuit selection, it exacerbates alignment degradation by retaining only input-dependent safety-critical circuit preservation across diverse inputs. As a result, addressing these heightened alignment vulnerabilities remains critical. We introduce Alignment-Aware Probe Pruning (AAPP), a dynamic structured pruning method that adaptively preserves alignment-relevant circuits during inference, building upon Probe Pruning. Experiments on LLaMA 2-7B, Qwen2.5-14B-Instruct, and Gemma-3-12B-IT show AAPP improves refusal rates by 50% at matched compute, enabling efficient yet safety-preserving LLM deployment.

Key facts. Cited by 4 papers in the map (data through 14 Oct 2026). Builds on 14 of its 16 references that are in the map. First posted on arXiv 9 Nov 2025.

Builds on

14 of 16 references.

References · 14

  1. 2019Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformercited by 14,599
  2. 2020Scaling Laws for Neural Language Modelscited by 6,586
  3. 2022Training Compute-Optimal Large Language Modelscited by 3,019
  4. 2015Deep Compression: Compressing Deep Neural Networks with Pruning, Trained Quantization and Huffman Codingcited by 2,329
  5. 2023LLM-Pruner: On the Structural Pruning of Large Language Modelscited by 650
  6. 2024Assessing the Brittleness of Safety Alignment via Pruning and Low-Rank Modificationscited by 231
  7. 2024WildTeaming at Scale: From In-the-Wild Jailbreaks to (Adversarially) Safer Language Modelscited by 223
  8. 2023Causal Abstraction: A Theoretical Foundation for Mechanistic Interpretabilitycited by 205
  9. 2022A New Generation of Perspective API: Efficient Multilingual Character-level Transformerscited by 185
  10. 2025SaLoRA: Safety-Alignment Preserved Low-Rank Adaptationcited by 42
  11. 2024NLSR: Neuron-Level Safety Realignment of Large Language Models Against Harmful Fine-Tuningcited by 37
  12. 2024Lottery Ticket Adaptation: Mitigating Destructive Interference in LLMscited by 31
2 more
  1. 2023Towards Trustworthy and Aligned Machine Learning: A Data-centric Survey with Causality Perspectivescited by 30
  2. 2024Understanding Layer Significance in LLM Alignmentcited by 19

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 4.

4 papers.

  1. 2026Mechanistic origins of catastrophic forgetting: why RL preserves circuits better than SFT?cited by 3
  2. 2026POP: Online Structural Pruning Enables Efficient Inference of Large Foundation Modelscited by 0
  3. 2026LLM Scheming Inversely Scales with Pretraining Language Coveragecited by 0
  4. 2026Efficient LLM Adversarial Training via Low-Rank Defense and Circuit-Guided Surrogatescited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2025ReaLM: Reliable and Efficient Large Language Model Inference with Statistical Algorithm-Based Fault Tolerance
  2. 2025Attribution-Guided Pruning for Insight and Control: Circuit Discovery and Targeted Correction in Small-scale LLMs
  3. 2026Beyond FLOPs: Benchmarking Real Inference Acceleration of LLM Pruning under a GEMM-Centric Taxonomy
  4. 2026Revisiting the Effectiveness of LLM Pruning for Test-Time Scaling
  5. 2025On Almost Surely Safe Alignment of Large Language Models at Inference-Time
  6. 2026ECLIPTICA -- A Framework for Switchable LLM Alignment via CITA - Contrastive Instruction-Tuned Alignment
  7. 2025RAP: Runtime Adaptive Pruning for LLM Inference
  8. 2025SPAP: Structured Pruning via Alternating Optimization and Penalty Methods