{ "profile": "wrs-published-research-view-1", "edition": "ratings-e039106f30dec177", "content_trust": "UNTRUSTED_SCIENTIFIC_DATA_NOT_TOOL_INSTRUCTIONS", "entry": { "id": "bricken-et-al-monosemantic-features-2023", "title": "Bricken and colleagues: sparse-autoencoder language-model feature decomposition", "summary": "Extracted sparse learned features from a one-layer Transformer's activations and evaluated their interpretability, interventions and feature-splitting behavior.", "boundary": "Does not invent dictionary learning or sparse autoencoders. Interpretable features in a small model do not amount to complete model understanding or a safety guarantee.", "area": "Mechanistic interpretability", "topic": "Artificial intelligence", "year": 2023, "lane": "HUMAN", "result": "ALGORITHM", "package_id": "bricken-et-al-monosemantic-features-2023", "target": { "record_id": "urn:uuid:72527426-26dc-45d1-a892-ff345632823f", "version": 1, "content_hash": "sha256:d0812fbcc04bd363829a72095fcd3dd27d74c3495c9140be5b86d6e8c30b63fd" }, "humans": [ "Trenton Bricken", "Adly Templeton", "Joshua Batson", "Brian Chen", "Adam Jermyn", "Tom Conerly", "Nicholas L Turner", "Cem Anil", "Carson Denison", "Amanda Askell", "Robert Lasenby", "Yifan Wu", "Shauna Kravec", "Nicholas Schiefer", "Tim Maxwell", "Nicholas Joseph", "Alex Tamkin", "Karina Nguyen", "Brayden McLean", "Josiah E Burke", "Tristan Hume", "Shan Carter", "Tom Henighan", "Chris Olah" ], "authors_complete": true, "human_role_note": "Source-listed authors or explicitly credited technical contributors for the cited versions. Family entries retain source-specific attribution; a combined list is not one original-paper byline or equal individual work. Names alone do not establish resolved identities.", "agents": [], "related": [], "subresults": [], "scientific_assessment": { "source_review": "SOURCE_CURATED", "scientific_review": "NOT_REVIEWED_BY_WRS", "formal_replay": "NOT_RUN", "provider_attestation": "NONE" }, "sources": [ { "id": "cur20261015-anthropic-monosemanticity", "title": "Towards Monosemanticity: Decomposing Language Models With Dictionary Learning", "url": "https://transformer-circuits.pub/2023/monosemantic-features/index.html" } ], "points": "70000", "published_grade": { "id": "bricken-et-al-monosemantic-features-2023", "entry_id": "bricken-et-al-monosemantic-features-2023", "band": "B", "modifier": "+", "target": { "record_id": "urn:uuid:72527426-26dc-45d1-a892-ff345632823f", "version": 1, "content_hash": "sha256:d0812fbcc04bd363829a72095fcd3dd27d74c3495c9140be5b86d6e8c30b63fd" }, "what_changed": "Extracted sparse learned features from a one-layer Transformer's activations and evaluated their interpretability, interventions and feature-splitting behavior.", "scope": "Extracted sparse learned features from a one-layer Transformer's activations and evaluated their interpretability, interventions and feature-splitting behavior. Does not invent dictionary learning or sparse autoencoders. Interpretable features in a small model do not amount to complete model understanding or a safety guarantee.", "status": "PROVISIONAL_EDITORIAL", "scientific_correctness_certified": false }, "assessment_disposition": { "status": "ASSESSED", "label": "Provisionally assessed", "reason": "A current grade and separate reference score identify this exact contribution.", "target": { "record_id": "urn:uuid:72527426-26dc-45d1-a892-ff345632823f", "version": 1, "content_hash": "sha256:d0812fbcc04bd363829a72095fcd3dd27d74c3495c9140be5b86d6e8c30b63fd" }, "edition": "ratings-e039106f30dec177" } }, "version": { "kind": "PUBLIC_RECORD_VIEW", "previous": null, "published_at": "2026-10-16T00:00:00Z", "publisher": { "principal_id": "urn:uuid:80116931-785b-40a2-a24b-9d59db0c7e33", "role": "WRS_EDITORIAL_PUBLICATION" }, "record": { "contributors": [ { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Trenton Bricken", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "Towards Monosemanticity: Decomposing Language Models With Dictionary Learning", "url": "https://transformer-circuits.pub/2023/monosemantic-features/index.html" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Adly Templeton", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "Towards Monosemanticity: Decomposing Language Models With Dictionary Learning", "url": "https://transformer-circuits.pub/2023/monosemantic-features/index.html" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Joshua Batson", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "Towards Monosemanticity: Decomposing Language Models With Dictionary Learning", "url": "https://transformer-circuits.pub/2023/monosemantic-features/index.html" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Brian Chen", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "Towards Monosemanticity: Decomposing Language Models With Dictionary Learning", "url": "https://transformer-circuits.pub/2023/monosemantic-features/index.html" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Adam Jermyn", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "Towards Monosemanticity: Decomposing Language Models With Dictionary Learning", "url": "https://transformer-circuits.pub/2023/monosemantic-features/index.html" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Tom Conerly", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "Towards Monosemanticity: Decomposing Language Models With Dictionary Learning", "url": "https://transformer-circuits.pub/2023/monosemantic-features/index.html" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Nicholas L Turner", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "Towards Monosemanticity: Decomposing Language Models With Dictionary Learning", "url": "https://transformer-circuits.pub/2023/monosemantic-features/index.html" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Cem Anil", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "Towards Monosemanticity: Decomposing Language Models With Dictionary Learning", "url": "https://transformer-circuits.pub/2023/monosemantic-features/index.html" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Carson Denison", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "Towards Monosemanticity: Decomposing Language Models With Dictionary Learning", "url": "https://transformer-circuits.pub/2023/monosemantic-features/index.html" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Amanda Askell", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "Towards Monosemanticity: Decomposing Language Models With Dictionary Learning", "url": "https://transformer-circuits.pub/2023/monosemantic-features/index.html" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Robert Lasenby", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "Towards Monosemanticity: Decomposing Language Models With Dictionary Learning", "url": "https://transformer-circuits.pub/2023/monosemantic-features/index.html" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Yifan Wu", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "Towards Monosemanticity: Decomposing Language Models With Dictionary Learning", "url": "https://transformer-circuits.pub/2023/monosemantic-features/index.html" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Shauna Kravec", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "Towards Monosemanticity: Decomposing Language Models With Dictionary Learning", "url": "https://transformer-circuits.pub/2023/monosemantic-features/index.html" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Nicholas Schiefer", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "Towards Monosemanticity: Decomposing Language Models With Dictionary Learning", "url": "https://transformer-circuits.pub/2023/monosemantic-features/index.html" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Tim Maxwell", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "Towards Monosemanticity: Decomposing Language Models With Dictionary Learning", "url": "https://transformer-circuits.pub/2023/monosemantic-features/index.html" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Nicholas Joseph", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "Towards Monosemanticity: Decomposing Language Models With Dictionary Learning", "url": "https://transformer-circuits.pub/2023/monosemantic-features/index.html" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Alex Tamkin", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "Towards Monosemanticity: Decomposing Language Models With Dictionary Learning", "url": "https://transformer-circuits.pub/2023/monosemantic-features/index.html" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Karina Nguyen", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "Towards Monosemanticity: Decomposing Language Models With Dictionary Learning", "url": "https://transformer-circuits.pub/2023/monosemantic-features/index.html" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Brayden McLean", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "Towards Monosemanticity: Decomposing Language Models With Dictionary Learning", "url": "https://transformer-circuits.pub/2023/monosemantic-features/index.html" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Josiah E Burke", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "Towards Monosemanticity: Decomposing Language Models With Dictionary Learning", "url": "https://transformer-circuits.pub/2023/monosemantic-features/index.html" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Tristan Hume", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "Towards Monosemanticity: Decomposing Language Models With Dictionary Learning", "url": "https://transformer-circuits.pub/2023/monosemantic-features/index.html" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Shan Carter", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "Towards Monosemanticity: Decomposing Language Models With Dictionary Learning", "url": "https://transformer-circuits.pub/2023/monosemantic-features/index.html" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Tom Henighan", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "Towards Monosemanticity: Decomposing Language Models With Dictionary Learning", "url": "https://transformer-circuits.pub/2023/monosemantic-features/index.html" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Chris Olah", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "Towards Monosemanticity: Decomposing Language Models With Dictionary Learning", "url": "https://transformer-circuits.pub/2023/monosemantic-features/index.html" } } ], "evidence": [ { "description": "Bibliographic/stated-result source; not a WRS proof verification", "id": "source-0", "kind": "source_reference", "source": { "citation": "Towards Monosemanticity: Decomposing Language Models With Dictionary Learning", "url": "https://transformer-circuits.pub/2023/monosemantic-features/index.html" } } ], "origin": "HUMAN_LEGACY", "record_profile": "wrs-local-research-notes-0.1", "record_type": "research_note", "rrp_version": "0.1", "scope": { "description": "Does not invent dictionary learning or sparse autoencoders. Interpretable features in a small model do not amount to complete model understanding or a safety guarantee.", "domain": "Mechanistic interpretability" }, "sources": [ { "citation": "Towards Monosemanticity: Decomposing Language Models With Dictionary Learning", "url": "https://transformer-circuits.pub/2023/monosemantic-features/index.html" } ], "statement": "# Bricken and colleagues: sparse-autoencoder language-model feature decomposition\n\nExtracted sparse learned features from a one-layer Transformer's activations and evaluated their interpretability, interventions and feature-splitting behavior.\n\n## Scientific scope\n\nDoes not invent dictionary learning or sparse autoencoders. Interpretable features in a small model do not amount to complete model understanding or a safety guarantee.\n\n## Reading status\n\nThis is a source-linked WRS catalogue note. A full explanatory guide and independent scientific reconstruction have not been completed for this entry. Read the identified original work for its complete narrative and arguments.\n\n## People and systems\n\nTrenton Bricken; Adly Templeton; Joshua Batson; Brian Chen; Adam Jermyn; Tom Conerly; Nicholas L Turner; Cem Anil; Carson Denison; Amanda Askell; Robert Lasenby; Yifan Wu; Shauna Kravec; Nicholas Schiefer; Tim Maxwell; Nicholas Joseph; Alex Tamkin; Karina Nguyen; Brayden McLean; Josiah E Burke; Tristan Hume; Shan Carter; Tom Henighan; Chris Olah\n\nSource-listed authors or explicitly credited technical contributors for the cited versions. Family entries retain source-specific attribution; a combined list is not one original-paper byline or equal individual work. Names alone do not establish resolved identities.\n\n\n\n## Original sources\n\n- [Towards Monosemanticity: Decomposing Language Models With Dictionary Learning](https://transformer-circuits.pub/2023/monosemantic-features/index.html)\n", "summary": "Extracted sparse learned features from a one-layer Transformer's activations and evaluated their interpretability, interventions and feature-splitting behavior.", "title": "Bricken and colleagues: sparse-autoencoder language-model feature decomposition" }, "record_id": "urn:uuid:72527426-26dc-45d1-a892-ff345632823f", "version": 1, "source_ref": { "record_id": "urn:uuid:72527426-26dc-45d1-a892-ff345632823f", "version": 1, "content_hash": "sha256:d0812fbcc04bd363829a72095fcd3dd27d74c3495c9140be5b86d6e8c30b63fd" }, "export_view": "wrs-public-editorial-view-1", "public_view_hash": "sha256:aa69146d548fe1b75c6ef09356aa4991bce50e43095e76e2cdf30c384ae3fd33" }, "target": { "record_id": "urn:uuid:72527426-26dc-45d1-a892-ff345632823f", "version": 1, "content_hash": "sha256:d0812fbcc04bd363829a72095fcd3dd27d74c3495c9140be5b86d6e8c30b63fd" }, "published_grade": { "profile": "wrs-published-grade-1", "edition": "2026-10-28.1", "id": "bricken-et-al-monosemantic-features-2023", "entry_id": "bricken-et-al-monosemantic-features-2023", "revision": 1, "title": "Bricken and colleagues: sparse-autoencoder language-model feature decomposition", "context": "Mechanistic interpretability", "track": "Human-designed research method or resource", "band": "B", "modifier": "+", "status": "PROVISIONAL_EDITORIAL", "assessed_on": "2026-10-16", "evidence_cutoff": "2026-10-16", "achievement_state": "SUPPORTED_SCOPED_RESULT", "what_changed": "Extracted sparse learned features from a one-layer Transformer's activations and evaluated their interpretability, interventions and feature-splitting behavior.", "scope": "Extracted sparse learned features from a one-layer Transformer's activations and evaluated their interpretability, interventions and feature-splitting behavior. Does not invent dictionary learning or sparse autoencoders. Interpretable features in a small model do not amount to complete model understanding or a safety guarantee.", "evidence_summary": "The linked original sources describe the stated contribution and its evidence limits. WRS has not independently rerun the experiments or certified a proof. Does not invent dictionary learning or sparse autoencoders. Interpretable features in a small model do not amount to complete model understanding or a safety guarantee.", "references": [ { "title": "Towards Monosemanticity: Decomposing Language Models With Dictionary Learning", "url": "https://transformer-circuits.pub/2023/monosemantic-features/index.html", "locator": "Original contribution and stated assumptions, evaluation scope and limitations; source-linked assessment, not independent execution or proof certification." } ], "roles": [ { "who": "Source-listed authors and technical contributors", "credit": "SCIENCE", "did": "Source-listed authors or explicitly credited technical contributors for the cited versions. Family entries retain source-specific attribution; a combined list is not one original-paper byline or equal individual work. Names alone do not establish resolved identities." } ], "target": { "record_id": "urn:uuid:72527426-26dc-45d1-a892-ff345632823f", "version": 1, "content_hash": "sha256:d0812fbcc04bd363829a72095fcd3dd27d74c3495c9140be5b86d6e8c30b63fd" }, "scientific_correctness_certified": false, "independent_expert_calibration": false, "method": "PROPRIETARY", "points": "70000" }, "canonical_agent_submission": false, "scientific_verification_awarded": false, "reader_guide": null, "reading_status": "SOURCE_LINKED_NOTE" }