BlackboxNLP-2025 MIB Shared Task: Improving Circuit Faithfulness via Better Edge Selection

TL;DR

This paper introduces improved methods for circuit discovery in mechanistic interpretability, using bootstrapping, ratio-based edge selection, and ILP to enhance circuit faithfulness across models.

Contribution

It presents three novel techniques—bootstrapping, ratio-based edge prioritization, and ILP formulation—that improve circuit discovery in the MIB benchmark.

Findings

More faithful circuits achieved

Outperforms prior approaches on multiple tasks

Code available for reproducibility

Abstract

One of the main challenges in mechanistic interpretability is circuit discovery, determining which parts of a model perform a given task. We build on the Mechanistic Interpretability Benchmark (MIB) and propose three key improvements to circuit discovery. First, we use bootstrapping to identify edges with consistent attribution scores. Second, we introduce a simple ratio-based selection strategy to prioritize strong positive-scoring edges, balancing performance and faithfulness. Third, we replace the standard greedy selection with an integer linear programming formulation. Our methods yield more faithful circuits and outperform prior approaches across multiple MIB tasks and models. Our code is available at: https://github.com/technion-cs-nlp/MIB-Shared-Task.