{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,4,17]],"date-time":"2026-04-17T19:19:33Z","timestamp":1776453573240,"version":"3.51.2"},"reference-count":0,"publisher":"Association for the Advancement of Artificial Intelligence (AAAI)","issue":"1","content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["AIES"],"abstract":"<jats:p>Mechanistic Interpretability aims to understand neural net-\nworks through causal explanations. We argue for the\nExplanatory View Hypothesis: that Mechanistic\nInterpretability re- search is a principled approach to\nunderstanding models be- cause neural networks contain\nimplicit explanations which can be extracted and\nunderstood. We hence show that Explanatory Faithfulness, an\nassessment of how well an explanation fits a model, is\nwell-defined. We propose a definition of Mechanistic\nInterpretability (MI) as the practice of producing\nModel-level, Ontic, Causal-Mechanistic, and Falsifiable\nexplanations of neural networks, allowing us to distinguish\nMI from other interpretability paradigms and detail MI\u2019s\ninherent limits. We formulate the Principle of Explanatory\nOptimism, a conjecture which we argue is a necessary\nprecondition for the success of Mechanistic\nInterpretability.<\/jats:p>","DOI":"10.1609\/aies.v8i1.36547","type":"journal-article","created":{"date-parts":[[2025,10,15]],"date-time":"2025-10-15T13:17:02Z","timestamp":1760534222000},"page":"265-278","source":"Crossref","is-referenced-by-count":1,"title":["A Mathematical Philosophy of Explanations in Mechanistic Interpretability"],"prefix":"10.1609","volume":"8","author":[{"given":"Kola","family":"Ayonrinde","sequence":"first","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Louis","family":"Jaburi","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"9382","published-online":{"date-parts":[[2025,10,15]]},"container-title":["Proceedings of the AAAI\/ACM Conference on AI, Ethics, and Society"],"original-title":[],"link":[{"URL":"https:\/\/ojs.aaai.org\/index.php\/AIES\/article\/download\/36547\/38685","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/ojs.aaai.org\/index.php\/AIES\/article\/download\/36547\/38685","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,10,15]],"date-time":"2025-10-15T13:17:03Z","timestamp":1760534223000},"score":1,"resource":{"primary":{"URL":"https:\/\/ojs.aaai.org\/index.php\/AIES\/article\/view\/36547"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,10,15]]},"references-count":0,"journal-issue":{"issue":"1","published-online":{"date-parts":[[2025,10,15]]}},"URL":"https:\/\/doi.org\/10.1609\/aies.v8i1.36547","relation":{},"ISSN":["3065-8365"],"issn-type":[{"value":"3065-8365","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,10,15]]}}}