{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:MMWQK7LXAJOO6U2VFOGM76Q27K","short_pith_number":"pith:MMWQK7LX","schema_version":"1.0","canonical_sha256":"632d057d77025cef53552b8ccffa1afa9c90ff9fed72a80d90f8f29852818a71","source":{"kind":"arxiv","id":"2406.15109","version":1},"attestation_state":"computed","paper":{"title":"Brain-Like Language Processing via a Shallow Untrained Multihead Attention Network","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Antoine Bosselut, Badr AlKhamissi, Greta Tuckute, Martin Schrimpf","submitted_at":"2024-06-21T12:54:03Z","abstract_excerpt":"Large Language Models (LLMs) have been shown to be effective models of the human language system, with some models predicting most explainable variance of brain activity in current datasets. Even in untrained models, the representations induced by architectural priors can exhibit reasonable alignment to brain data. In this work, we investigate the key architectural components driving the surprising alignment of untrained models. To estimate LLM-to-brain similarity, we first select language-selective units within an LLM, similar to how neuroscientists identify the language network in the human "},"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":"2406.15109","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2024-06-21T12:54:03Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"f9959b818d55d5cb9040c1ab5706755f0205751bd844d541a63e1a6903870e64","abstract_canon_sha256":"17d31387f527d4ba1f106ad87da7249bc56d9cfa6e72e0904406b2387f7d6c51"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:35:14.177457Z","signature_b64":"miZ42tOnQ5R0ftHQ/0x/ZReNI07hWJY8aV/e095lwB6R/XYdFkns6oojvb59oPpwUXXei4MrirBc4YpiVRLdBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"632d057d77025cef53552b8ccffa1afa9c90ff9fed72a80d90f8f29852818a71","last_reissued_at":"2026-07-05T08:35:14.177009Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:35:14.177009Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Brain-Like Language Processing via a Shallow Untrained Multihead Attention Network","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Antoine Bosselut, Badr AlKhamissi, Greta Tuckute, Martin Schrimpf","submitted_at":"2024-06-21T12:54:03Z","abstract_excerpt":"Large Language Models (LLMs) have been shown to be effective models of the human language system, with some models predicting most explainable variance of brain activity in current datasets. Even in untrained models, the representations induced by architectural priors can exhibit reasonable alignment to brain data. In this work, we investigate the key architectural components driving the surprising alignment of untrained models. To estimate LLM-to-brain similarity, we first select language-selective units within an LLM, similar to how neuroscientists identify the language network in the human "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2406.15109","kind":"arxiv","version":1},"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/2406.15109/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":"2406.15109","created_at":"2026-07-05T08:35:14.177068+00:00"},{"alias_kind":"arxiv_version","alias_value":"2406.15109v1","created_at":"2026-07-05T08:35:14.177068+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2406.15109","created_at":"2026-07-05T08:35:14.177068+00:00"},{"alias_kind":"pith_short_12","alias_value":"MMWQK7LXAJOO","created_at":"2026-07-05T08:35:14.177068+00:00"},{"alias_kind":"pith_short_16","alias_value":"MMWQK7LXAJOO6U2V","created_at":"2026-07-05T08:35:14.177068+00:00"},{"alias_kind":"pith_short_8","alias_value":"MMWQK7LX","created_at":"2026-07-05T08:35:14.177068+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2504.21047","citing_title":"Model Connectomes: A Generational Approach to Data-Efficient Language Models","ref_index":46,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/MMWQK7LXAJOO6U2VFOGM76Q27K","json":"https://pith.science/pith/MMWQK7LXAJOO6U2VFOGM76Q27K.json","graph_json":"https://pith.science/api/pith-number/MMWQK7LXAJOO6U2VFOGM76Q27K/graph.json","events_json":"https://pith.science/api/pith-number/MMWQK7LXAJOO6U2VFOGM76Q27K/events.json","paper":"https://pith.science/paper/MMWQK7LX"},"agent_actions":{"view_html":"https://pith.science/pith/MMWQK7LXAJOO6U2VFOGM76Q27K","download_json":"https://pith.science/pith/MMWQK7LXAJOO6U2VFOGM76Q27K.json","view_paper":"https://pith.science/paper/MMWQK7LX","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2406.15109&json=true","fetch_graph":"https://pith.science/api/pith-number/MMWQK7LXAJOO6U2VFOGM76Q27K/graph.json","fetch_events":"https://pith.science/api/pith-number/MMWQK7LXAJOO6U2VFOGM76Q27K/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/MMWQK7LXAJOO6U2VFOGM76Q27K/action/timestamp_anchor","attest_storage":"https://pith.science/pith/MMWQK7LXAJOO6U2VFOGM76Q27K/action/storage_attestation","attest_author":"https://pith.science/pith/MMWQK7LXAJOO6U2VFOGM76Q27K/action/author_attestation","sign_citation":"https://pith.science/pith/MMWQK7LXAJOO6U2VFOGM76Q27K/action/citation_signature","submit_replication":"https://pith.science/pith/MMWQK7LXAJOO6U2VFOGM76Q27K/action/replication_record"}},"created_at":"2026-07-05T08:35:14.177068+00:00","updated_at":"2026-07-05T08:35:14.177068+00:00"}