{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:AHFY3J72PVWQR4OEOXOPAW5FNA","short_pith_number":"pith:AHFY3J72","schema_version":"1.0","canonical_sha256":"01cb8da7fa7d6d08f1c475dcf05ba56800372fbc821843e92ac5e620141dde6b","source":{"kind":"arxiv","id":"2504.19475","version":3},"attestation_state":"computed","paper":{"title":"Prisma: An Open Source Toolkit for Mechanistic Interpretability in Vision and Video","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CV","authors_text":"Blake Aaron Richards, Danilo Bzdok, Edward Stevinson, Lee Sharkey, Lorenz Hufe, Praneet Suresh, Robert Graham, Sebastian Lapuschkin, Sonia Joseph, Yash Vadi","submitted_at":"2025-04-28T04:31:24Z","abstract_excerpt":"Robust tooling and publicly available pre-trained models have helped drive recent advances in mechanistic interpretability for language models. However, similar progress in vision mechanistic interpretability has been hindered by the lack of accessible frameworks and pre-trained weights. We present Prisma (Access the codebase here: https://github.com/Prisma-Multimodal/ViT-Prisma), an open-source framework designed to accelerate vision mechanistic interpretability research, providing a unified toolkit for accessing 75+ vision and video transformers; support for sparse autoencoder (SAE), transco"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2504.19475","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2025-04-28T04:31:24Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"1aba015016389f8a6c7690cae88c493ec0dcbdd8834b9bb0ac42b9c6e432ca19","abstract_canon_sha256":"ec4a2eb28078ca873a8a882eff503ed9baa3d79ae087a781168f76f592325125"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:14:58.782881Z","signature_b64":"R8E/1cinAjf6GlrOIoldD9WtwN0/7qsLjN6hOBlj3qUnbZMHCfo/JoZLwVtqbbgzZcny9ujyzchwnSgZl6CWBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"01cb8da7fa7d6d08f1c475dcf05ba56800372fbc821843e92ac5e620141dde6b","last_reissued_at":"2026-07-05T11:14:58.782228Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:14:58.782228Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Prisma: An Open Source Toolkit for Mechanistic Interpretability in Vision and Video","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CV","authors_text":"Blake Aaron Richards, Danilo Bzdok, Edward Stevinson, Lee Sharkey, Lorenz Hufe, Praneet Suresh, Robert Graham, Sebastian Lapuschkin, Sonia Joseph, Yash Vadi","submitted_at":"2025-04-28T04:31:24Z","abstract_excerpt":"Robust tooling and publicly available pre-trained models have helped drive recent advances in mechanistic interpretability for language models. However, similar progress in vision mechanistic interpretability has been hindered by the lack of accessible frameworks and pre-trained weights. We present Prisma (Access the codebase here: https://github.com/Prisma-Multimodal/ViT-Prisma), an open-source framework designed to accelerate vision mechanistic interpretability research, providing a unified toolkit for accessing 75+ vision and video transformers; support for sparse autoencoder (SAE), transco"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2504.19475","kind":"arxiv","version":3},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2504.19475/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"2504.19475","created_at":"2026-07-05T11:14:58.782307+00:00"},{"alias_kind":"arxiv_version","alias_value":"2504.19475v3","created_at":"2026-07-05T11:14:58.782307+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2504.19475","created_at":"2026-07-05T11:14:58.782307+00:00"},{"alias_kind":"pith_short_12","alias_value":"AHFY3J72PVWQ","created_at":"2026-07-05T11:14:58.782307+00:00"},{"alias_kind":"pith_short_16","alias_value":"AHFY3J72PVWQR4OE","created_at":"2026-07-05T11:14:58.782307+00:00"},{"alias_kind":"pith_short_8","alias_value":"AHFY3J72","created_at":"2026-07-05T11:14:58.782307+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":7,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.16196","citing_title":"When Confidence Lacks Concepts: Interpretable OOD Detection via Representation Perturbations","ref_index":15,"is_internal_anchor":false},{"citing_arxiv_id":"2606.10080","citing_title":"VFUSE: Virulent Feature Understanding with Sparse autoEncoders","ref_index":3,"is_internal_anchor":false},{"citing_arxiv_id":"2606.28399","citing_title":"Meta-learning as a principle for human-like visual representations","ref_index":49,"is_internal_anchor":false},{"citing_arxiv_id":"2606.26396","citing_title":"At the Edge of Understanding: Sparse Autoencoders Trace The Limits of Transformer Generalization","ref_index":10,"is_internal_anchor":false},{"citing_arxiv_id":"2510.18900","citing_title":"Foundation Models for Discovery and Exploration in Chemical Space","ref_index":49,"is_internal_anchor":false},{"citing_arxiv_id":"2604.22477","citing_title":"Contrastive Semantic Projection: Faithful Neuron Labeling with Contrastive Examples","ref_index":14,"is_internal_anchor":false},{"citing_arxiv_id":"2604.14477","citing_title":"Seeing Through Circuits: Faithful Mechanistic Interpretability for Vision Transformers","ref_index":21,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/AHFY3J72PVWQR4OEOXOPAW5FNA","json":"https://pith.science/pith/AHFY3J72PVWQR4OEOXOPAW5FNA.json","graph_json":"https://pith.science/api/pith-number/AHFY3J72PVWQR4OEOXOPAW5FNA/graph.json","events_json":"https://pith.science/api/pith-number/AHFY3J72PVWQR4OEOXOPAW5FNA/events.json","paper":"https://pith.science/paper/AHFY3J72"},"agent_actions":{"view_html":"https://pith.science/pith/AHFY3J72PVWQR4OEOXOPAW5FNA","download_json":"https://pith.science/pith/AHFY3J72PVWQR4OEOXOPAW5FNA.json","view_paper":"https://pith.science/paper/AHFY3J72","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2504.19475&json=true","fetch_graph":"https://pith.science/api/pith-number/AHFY3J72PVWQR4OEOXOPAW5FNA/graph.json","fetch_events":"https://pith.science/api/pith-number/AHFY3J72PVWQR4OEOXOPAW5FNA/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/AHFY3J72PVWQR4OEOXOPAW5FNA/action/timestamp_anchor","attest_storage":"https://pith.science/pith/AHFY3J72PVWQR4OEOXOPAW5FNA/action/storage_attestation","attest_author":"https://pith.science/pith/AHFY3J72PVWQR4OEOXOPAW5FNA/action/author_attestation","sign_citation":"https://pith.science/pith/AHFY3J72PVWQR4OEOXOPAW5FNA/action/citation_signature","submit_replication":"https://pith.science/pith/AHFY3J72PVWQR4OEOXOPAW5FNA/action/replication_record"}},"created_at":"2026-07-05T11:14:58.782307+00:00","updated_at":"2026-07-05T11:14:58.782307+00:00"}