1
Demystifying Variance in Circuit Discovery of LLMs
揭秘LLM电路发现中的方差来源，为可解释性研究提供新视角
arXiv:2606.16920v1 Announce Type: cross Abstract: Circuit discovery is a key technique in mechanistic interpretability to pinpoint the model component…
2
Multi-Granular Node Pruning for Causal Circuit Discovery
新论文提出多粒度节点剪枝方法，高效发现因果电路，提升模型可解释性。
arXiv:2512.10903v2 Announce Type: replace Abstract: Circuit discovery aims to identify minimal subnetworks that are responsible for specific behaviors…
3
Revealing Algorithmic Deductive Circuits for Logical Reasoning
揭秘逻辑推理背后的算法演绎电路，为可解释AI提供新视角
arXiv:2605.27824v1 Announce Type: new Abstract: Recent studies have shown that Large Language Models (LLMs) can achieve strong reasoning performance b…