{"release":{"schemaVersion":"maha-epistemic-release/1.0","releaseId":"epirelease_a67e8b03cb744ab1b23ddc6acadf37a3","releaseKind":"initial","status":"active","recordId":"urn:maha:record:mechanistic-interpretability-toy-models-of-superposition","domainSlug":"mechanistic-interpretability","targetSha256":"sha256:a984a5310affae3ce9141323a019b36a57e03d4e87430b5d82591c391afc908e","canonicalPath":"/knowledge/mechanistic-interpretability/methods/mechanistic-interpretability-toy-models-of-superposition","canonicalVersion":"1.0.0","supersedesReleaseId":null,"approvals":[{"scope":"boundary-adequacy","reviewId":"epireview_55471f3eaae345e88c1b1b564cdfc484","reviewSha256":"sha256:98771f85cdbd040cea299d2ed4ee23b51088823b105189cb998e5e6bc5bb7e13","reviewedAt":"2026-08-25T09:34:37.055Z","reviewerKind":"internal-editorial","reviewMethod":"Deterministic exact-hash checks, public source-resolution evidence, source-to-claim audit, unsupported-inference audit, and AI-assisted internal editorial inspection. No external reviewer participated."},{"scope":"domain-fidelity","reviewId":"epireview_0e41fbc6c038428ebb9668e67a71e59b","reviewSha256":"sha256:f4cf5fbde8c0dad49f137c7e56d82db738d3b4cee5b929b688b082b9d649a981","reviewedAt":"2026-08-25T09:34:36.974Z","reviewerKind":"internal-editorial","reviewMethod":"Deterministic exact-hash checks, public source-resolution evidence, source-to-claim audit, unsupported-inference audit, and AI-assisted internal editorial inspection. No external reviewer participated."},{"scope":"rights-and-locator","reviewId":"epireview_d7614168e5e345d6b27ac17e705126a5","reviewSha256":"sha256:a4e2ed79a164a5b54d29642bed0da5bac89376196c15797753908584554ad96d","reviewedAt":"2026-08-25T09:34:37.149Z","reviewerKind":"internal-editorial","reviewMethod":"Deterministic exact-hash checks, public source-resolution evidence, source-to-claim audit, unsupported-inference audit, and AI-assisted internal editorial inspection. No external reviewer participated."},{"scope":"source-fidelity","reviewId":"epireview_7b712ffaa8354e859dfeba62a629b22d","reviewSha256":"sha256:d4dcfa3c489ceeeb53fc60d7fccb83e1ca4bc5516e2c558b4b7aeb6578c013f9","reviewedAt":"2026-08-25T09:34:36.727Z","reviewerKind":"internal-editorial","reviewMethod":"Deterministic exact-hash checks, public source-resolution evidence, source-to-claim audit, unsupported-inference audit, and AI-assisted internal editorial inspection. No external reviewer participated."}],"assuranceTier":"internally-reviewed-canonical","releaseAuthority":{"authoritySha256":"sha256:38616eafd13c4ad603a3a3a9e445f3862fb29b54385aa571c197a2923e302785","attribution":"withheld-by-consent"},"publicChangeSummary":"Initial canonical publication in the eight-domain internal-review throughput canary.","recordSha256":"sha256:65e9b289d4e37ce598e80f3ea2a4abdb947ed9685cc0f77ee4013a6855fcbc75","gateDecision":{"reasons":[],"recordId":"urn:maha:record:mechanistic-interpretability-toy-models-of-superposition","publicEligible":true,"evaluatedAgainst":"maha-epistemic/1.0"},"releasedAt":"2026-08-25T09:35:33.232Z","releaseSha256":"sha256:062dfe5d227625151665a84155b8dd8194df39b3f09f681b6ea625faae226db7","withdrawal":null},"provenance":{"schemaVersion":"maha-epistemic/1.0","evidencePolicyVersion":"mps/0.1","recordId":"urn:maha:record:mechanistic-interpretability-toy-models-of-superposition","canonicalPath":"/knowledge/mechanistic-interpretability/methods/mechanistic-interpretability-toy-models-of-superposition","contentHash":"sha256:65e9b289d4e37ce598e80f3ea2a4abdb947ed9685cc0f77ee4013a6855fcbc75","generatedAt":"2026-08-25T09:35:33.232Z","publicationDecision":{"recordId":"urn:maha:record:mechanistic-interpretability-toy-models-of-superposition","publicEligible":true,"evaluatedAgainst":"maha-epistemic/1.0","reasons":[]},"claims":[{"id":"urn:maha:claim:mechanistic-interpretability-toy-models-of-superposition","scope":"Limited to Definitions, toy models, geometry, sparsity, and feature-interference experiments. in “Toy Models of Superposition”; this candidate records the concept boundary and does not pool results from uncited systems or studies.","boundary":"Toy models of superposition does not by itself establish system-level performance, safety, manufacturability, scalability, economic advantage, clinical benefit, or deployment readiness.","claimKind":"theoretical-model","sourceIds":["source-mechanistic-interpretability-superposition"],"statement":"The cited source supports treating toy models of superposition as a distinct method within the stated mechanistic interpretability scope.","replication":{"asOfDate":"2026-08-24","assessment":"Independent replication and cross-platform transfer have not been compiled for this candidate; the evidence maturity refers only to the bounded source contract.","independentReplicationCount":null},"uncertainty":{"kind":"qualitative","statement":"No cross-source quantitative interval is asserted. Definitions, operating conditions, samples, instruments, and outcome measures must be checked against the exact cited locator during review."},"evidenceMaturity":"single-study"}],"sources":[{"id":"source-mechanistic-interpretability-superposition","url":"https://transformer-circuits.pub/2022/toy_model/index.html","title":"Toy Models of Superposition","rights":{"note":"The candidate uses original boundary language and a short paraphrase linked to the cited source. No source passage, figure, or table is reproduced.","basis":"citation-with-paraphrase","quotationUsed":false},"authors":["Nelson Elhage","Tristan Hume","Catherine Olsson","et al."],"boundary":"A toy-model mechanism does not establish that every feature in a production model has the same geometry or semantics.","publisher":"Transformer Circuits Thread","establishes":"The work develops toy models in which neural networks represent more features than available dimensions under specified sparsity conditions.","identifiers":[{"value":"https://transformer-circuits.pub/2022/toy_model/index.html","scheme":"url"}],"publishedAt":"2022-09-14","exactLocator":"Definitions, toy models, geometry, sparsity, and feature-interference experiments."}],"reviewEvents":[{"reviewId":"epireview_7b712ffaa8354e859dfeba62a629b22d","scope":"source-fidelity","targetSha256":"sha256:a984a5310affae3ce9141323a019b36a57e03d4e87430b5d82591c391afc908e","reviewedAt":"2026-08-25T09:34:36.727Z","verdict":"approve","supersedesReviewId":null},{"reviewId":"epireview_0e41fbc6c038428ebb9668e67a71e59b","scope":"domain-fidelity","targetSha256":"sha256:a984a5310affae3ce9141323a019b36a57e03d4e87430b5d82591c391afc908e","reviewedAt":"2026-08-25T09:34:36.974Z","verdict":"approve","supersedesReviewId":null},{"reviewId":"epireview_55471f3eaae345e88c1b1b564cdfc484","scope":"boundary-adequacy","targetSha256":"sha256:a984a5310affae3ce9141323a019b36a57e03d4e87430b5d82591c391afc908e","reviewedAt":"2026-08-25T09:34:37.055Z","verdict":"approve","supersedesReviewId":null},{"reviewId":"epireview_d7614168e5e345d6b27ac17e705126a5","scope":"rights-and-locator","targetSha256":"sha256:a984a5310affae3ce9141323a019b36a57e03d4e87430b5d82591c391afc908e","reviewedAt":"2026-08-25T09:34:37.149Z","verdict":"approve","supersedesReviewId":null}]},"privacyBoundary":"Operational actor fingerprints, bearer credentials, private reviewer profiles, affiliations, conflicts, and non-consented authority identity fields are excluded."}