Metraipolis › archive › dc_e256b94aedb2
MURANO: Design, Run, and Reproduce Mechanistic Interpretability Experiments as Composable Pipelines — arXiv:2608.30662
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2608.30662
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-17T08:07:40.927Z
- SHA-256 of the archived text
799559234aee46ff6ddea08a0fb16bb25eaf002bf93fd10ff05dcbf0cd89ce04- Stable URL of this page
https://metraipolis.com/documento/dc_e256b94aedb2
How to cite
MURANO: Design, Run, and Reproduce Mechanistic Interpretability Experiments as Composable Pipelines — arXiv:2608.30662 — arXiv (en), https://arxiv.org/abs/2608.30662 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_e256b94aedb2 (Metraipolis archive dc_e256b94aedb2, sha256 799559234aee46ff6ddea08a0fb16bb25eaf002bf93fd10ff05dcbf0cd89ce04)
BibTeX
@misc{dce256b94aedb2,
title = {MURANO: Design, Run, and Reproduce Mechanistic Interpretability Experiments as Composable Pipelines — arXiv:2608.30662},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2608.30662},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_e256b94aedb2; sha256 799559234aee46ff6ddea08a0fb16bb25eaf002bf93fd10ff05dcbf0cd89ce04},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.30662. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
This paper presents Murano, an open source framework for designing, running, and reproducing mechanistic interpretability studies of large language models, intended for researchers across disciplines. These studies often combine loading, recording, attribution, intervention, and evaluation, while existing libraries tend to focus on different parts of this workflow. As a result, researchers using several libraries may need to adapt outputs from one for use by another. To bridge this gap, Murano represents operations from these five areas as composable steps. Steps exchange named result artifacts and declare the inputs they require and the outputs they produce. A pipeline executes its steps in the order supplied, and Murano uses canonical addresses when component identities pass between operations. Murano builds on existing interpretability and machine learning libraries. We demonstrate Murano through two reproductions of established interpretability studies and one illustrative sparse autoencoder case study. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-01 in cs.CL). Categorie arXiv: Computation and Language (cs.CL). Pagina del lavoro: https://arxiv.org/abs/2608.30662 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-10-17T07:47:52.841Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-09-01?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).