paper-with-me

홈 › Papers

Towards Automated Circuit Discovery for Mechanistic Interpretability

2023-04-28 · NeurIPS 2023 11 · Arthur Conmy, Augustine N. Mavor-Parker, Aengus Lynch, Stefan Heimersheim, Adrià Garriga-Alonso

Through considerable effort and intuition, several recent works have reverse-engineered nontrivial behaviors of transformer models. This paper systematizes the mechanistic interpretability process they followed. First, researchers choose a metric and dataset that elicit the desired model behavior. Then, they apply activation patching to find which abstract neural network units are involved in the behavior. By varying the dataset, metric, and units under investigation, researchers can understand the functionality of each component. We automate one of the process' steps: to identify the circuit that implements the specified behavior in the model's computational graph. We propose several algorithms and reproduce previous interpretability results to validate them. For example, the ACDC algorithm rediscovered 5/5 of the component types in a circuit in GPT-2 Small that computes the Greater-Than operation. ACDC selected 68 of the 32,000 edges in GPT-2 Small, all of which were manually found by previous work. Our code is available at https://github.com/ArthurConmy/Automatic-Circuit-Discovery.

📄 PDF Abstract BibTeX arXiv:2304.14997

Code (4)

arthurconmy/automatic-circuit-discovery 공식 구현 pytorch
neelnanda-io/transformerlens 공식 구현 jax
FlyingPumba/circuits-benchmark jax
flyingpumba/interpbench jax

Methods 이 논문이 사용한 방법론

Multi-Head Attention 설명 없음
Attention 설명 없음
Linear Layer A Linear Layer is a projection $\mathbf{XW + b}$.
Cosine Annealing Cosine Annealing is a type of learning rate schedule that has the effect of starting with a large learning rate that is relatively rapidly decreased to a minimum value before…
Adam 설명 없음
Weight Decay 설명 없음
Refunds@Expedia|||How do I get a full refund from Expedia? “How do I get a full refund from Expedia? How do I get a full refund from Expedia? – Call ☎️ +1-(888) 829 (0881) or +1-805-330-4056 or +1-805-330-4056 for Quick Help &…
Residual Connection 설명 없음

Similar Papers 제목 키워드 기반

Evaluating Brain-Inspired Modular Training in Automated Circuit Discovery for Mechanistic Interpretability

2024-01-08 · Jatin Nainani

Large Language Models (LLMs) have experienced a rapid rise in AI, changing a wide range of applications with their advanced capabilities. As these models become increasingly integral to decision-making, the need for thor…

Decision Making

Formal Mechanistic Interpretability: Automated Circuit Discovery with Provable Guarantees

2026-02-18 · Itamar Hadad, Guy Katz, Shahaf Bassan arxiv

*Automated circuit discovery* is a central tool in mechanistic interpretability for identifying the internal components of neural networks responsible for specific behaviors. While prior methods have made significant pro…

MechRL: Reinforcement Learning Agents Perform Circuit Discovery for Mechanistic Interpretability

2026-05-25 · Barsat Khadka arxiv

Mechanistic interpretability seeks to explain a model's behaviour by finding its circuit: the sparse subgraph of the model's computation that is causally responsible for it. Automated methods have made this search system…

Reinforcement Learning

PAHQ: Accelerating Automated Circuit Discovery through Mixed-Precision Inference Optimization

2025-10-27 · Xinhai Wang, Shu Yang, Liangyu Wang, Lin Zhang 외 arxiv

Circuit discovery, which involves identifying sparse and task-relevant subnetworks in pre-trained language models, is a cornerstone of mechanistic interpretability. Automated Circuit Discovery (ACDC) has emerged as a piv…

Efficient Automated Circuit Discovery in Transformers using Contextual Decomposition

2024-07-01 · Aliyah R. Hsu, Georgia Zhou, Yeshwanth Cherapanamjeri, Yaxuan Huang 외

Automated mechanistic interpretation research has attracted great interest due to its potential to scale explanations of neural network internals to large models. Existing automated circuit discovery work relies on activ…

Computational EfficiencySST-2