paper-with-me

홈 › Papers

Distilled Circuits: A Mechanistic Study of Internal Restructuring in Knowledge Distillation

2025-05-16 · Reilly Haskins, Benjamin Adams

Knowledge distillation compresses a larger neural model (teacher) into smaller, faster student models by training them to match teacher outputs. However, the internal computational transformations that occur during this process remain poorly understood. We apply techniques from mechanistic interpretability to analyze how internal circuits, representations, and activation patterns differ between teacher and student. Focusing on GPT2-small and its distilled counterpart DistilGPT2, we find that student models reorganize, compress, and discard teacher components, often resulting in stronger reliance on fewer individual components. To quantify functional alignment beyond output similarity, we introduce an alignment metric based on influence-weighted component similarity, validated across multiple tasks. Our findings reveal that while knowledge distillation preserves broad functional behaviors, it also causes significant shifts in internal computation, with important implications for the robustness and generalization capacity of distilled models.

📄 PDF Abstract BibTeX arXiv:2505.10822

Code (1)

reih02/distilled_circuits 공식 구현

Tasks

Knowledge Distillation

Methods 이 논문이 사용한 방법론

Knowledge Distillation A very simple way to improve the performance of almost any machine learning algorithm is to train many different models on the same data and then to average their predictions.…

Similar Papers 제목 키워드 기반

Seeing Through Circuits: Faithful Mechanistic Interpretability for Vision Transformers

2026-04-15 · Nina Żukowska, Wolfgang Stammer, Bernt Schiele, Jonas Fischer arxiv

Transparency of neural networks' internal reasoning is at the heart of interpretability research, adding to trust, safety, and understanding of these models. The field of mechanistic interpretability has recently focused…

Restructuring Tractable Probabilistic Circuits

2024-11-19 · Honghua Zhang, Benjie Wang, Marcelo Arenas, Guy Van Den Broeck

Probabilistic circuits (PCs) is a unifying representation for probabilistic models that support tractable inference. Numerous applications of PCs like controllable text generation depend on the ability to efficiently mul…

Text Generation

On Mechanistic Circuits for Extractive Question-Answering

2025-02-12 · Samyadeep Basu, Vlad Morariu, Zichao Wang, Ryan Rossi 외

Large language models are increasingly used to process documents and facilitate question-answering on them. In our paper, we extract mechanistic circuits for this real-world language modeling task: context-augmented lang…

Extractive Question-AnsweringLanguage ModelingLanguage ModellingQuestion Answering

Wired for Overconfidence: A Mechanistic Perspective on Inflated Verbalized Confidence in LLMs

2026-04-01 · Tianyi Zhao, Yinhan He, Wendy Zheng, Yujie Zhang 외 arxiv

Large language models are often not just wrong, but \emph{confidently wrong}: when they produce factually incorrect answers, they tend to verbalize overly high confidence rather than signal uncertainty. Such verbalized o…

Triggers Hijack Language Circuits: A Mechanistic Analysis of Backdoor Behaviors in Large Language Models

2026-02-11 · Théo Lasnier, Wissam Antoun, Francis Kulumba, Djamé Seddah arxiv

Backdoor attacks pose significant security risks for Large Language Models (LLMs), yet the internal mechanisms by which triggers operate remain poorly understood. We present the first mechanistic analysis of language-swi…