arXiv AI August 26, 2026 By Nakyung Lee, Sangwoo Hong, Jungwoo Lee

Mechanistic Circuit Identification for Controllable Data Generation

The paper introduces a circuit‑grounded framework that links training‑dynamics‑based data valuation with mechanistic interpretability. It defines data quality along learnability, challenge, and alignment, identifies internal model circuits that control these utilities, and uses them as controllable interfaces for data generation. The authors present SAMS, a stage‑aware scheduling method that steers circuit‑guided data to match the model’s evolving optimization needs, achieving more diverse and effective data than prompt‑based baselines on multiple‑choice QA tasks.

Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv AI.

arXiv Machine Learning

Aug 13

Epiplexity Guided Data Selection and Generation for Out-of-Distribution Generalization

arXiv:2608. 11746v1 Announce Type: new Abstract: Modern systems are increasingly expected to transfer across tasks not specified during training.

By Ellen Su, Andres Potapczynski, Shikai Qiu, Edward Hughes, Andrew Gordon Wilson

fine-tuning

arXiv AI

Aug 26

Beyond Static Interpretability: Anticipating Post-SFT Mechanisms from Pre-SFT Parameters for Better Tuning

arXiv:2608.24482v1 Announce Type: cross Abstract: Mechanistic Localization bridges mechanistic interpretability and post-training optimization by isolating critical parameters via interpretative appr...

By Hang Chen, Jiaying Zhu, Wenya Wang

fine-tuning safety

arXiv AI

Sep 2

S^3martCirc: Self-supervised Smart Circuit Discovery

S^3martCirc is a self‑supervised framework that jointly discovers and interprets neural circuits in large language models, rather than treating circuit discovery and functional interpretation as separate stages. It abstracts node behavior into two general computational roles that generalize across tasks and introduces a quantitative metric for assigning these roles, enabling simultaneous identification of importance and function. Experiments demonstrate that S^3martCirc outperforms existing methods in circuit discovery.

By Wendy Zheng, Yinhan He, Liang Wu, Jundong Li

llms nlp safety

arXiv AI

Jun 9

Mechanistic Data Attribution: Tracing the Training Origins of Interpretable LLM Units

arXiv:2601. 21996v2 Announce Type: replace-cross Abstract: While Mechanistic Interpretability has identified interpretable circuits in LLMs, their causal origins in training data remain elusive.

By Jianhui Chen, Yuzhang Luo, Liangming Pan

llms safety

arXiv AI

Jun 24

Can Language Model Agents be Helpful Circuit Explainers in Mechanistic Interpretability?

arXiv:2606. 24026v1 Announce Type: new Abstract: Mechanistic interpretability has made substantial progress in automatically localizing circuits, but explaining what localized components do remains labor-intensive and difficult to standardize.

By Ayan Antik Khan, Harsh Kohli, Yuekun Yao, Huan Sun, Ziyu Yao

llms agents benchmarks safety

Hugging Face Trending Papers

Aug 25

Beyond Static Interpretability: Anticipating Post-SFT Mechanisms from Pre-SFT Parameters for Better Tuning

Mechanistic Localization bridges mechanistic interpretability and post-training optimization by isolating critical parameters via interpretative approaches and then guiding parameter-efficient Supervi...

fine-tuning safety

One email a morning, machine-written

Related stories

Epiplexity Guided Data Selection and Generation for Out-of-Distribution Generalization

Beyond Static Interpretability: Anticipating Post-SFT Mechanisms from Pre-SFT Parameters for Better Tuning

S^3martCirc: Self-supervised Smart Circuit Discovery

Mechanistic Data Attribution: Tracing the Training Origins of Interpretable LLM Units

Can Language Model Agents be Helpful Circuit Explainers in Mechanistic Interpretability?

Beyond Static Interpretability: Anticipating Post-SFT Mechanisms from Pre-SFT Parameters for Better Tuning