We are pleased to announce that our recent paper Many Circuits, One Mechanism: Input Variation and Evaluation Granularity in Circuit Discovery has received the Featured Certification from the Transactions on Machine Learning Research (TMLR).

The Featured Certification recognizes papers that demonstrate particularly strong quality, originality, and significance, and is awarded to less than 10% of published TMLR papers.

Congratulations to the authors Alireza Bayat Makou, Jingcheng Niu, Subhabrata Dutta, Iryna Gurevych on this achievement!