{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:XDT2HBVJKKHT42UDKSDZDBIO2C","short_pith_number":"pith:XDT2HBVJ","schema_version":"1.0","canonical_sha256":"b8e7a386a9528f3e6a83548791850ed09769619238e0731158f5ba5b048d8368","source":{"kind":"arxiv","id":"2402.15055","version":2},"attestation_state":"computed","paper":{"title":"Interpreting Context Look-ups in Transformers: Investigating Attention-MLP Interactions","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Clement Neo, Fazl Barez, Shay B. Cohen","submitted_at":"2024-02-23T02:15:47Z","abstract_excerpt":"Understanding the inner workings of large language models (LLMs) is crucial for advancing their theoretical foundations and real-world applications. While the attention mechanism and multi-layer perceptrons (MLPs) have been studied independently, their interactions remain largely unexplored. This study investigates how attention heads and next-token neurons interact in LLMs to predict new words. We propose a methodology to identify next-token neurons, find prompts that highly activate them, and determine the upstream attention heads responsible. We then generate and evaluate explanations for t"},"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":"2402.15055","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-02-23T02:15:47Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"54ad6308aa9e31aa17af89a6018e1849d7d5c0935d7319ec9eb4b6b6aa8d20a0","abstract_canon_sha256":"fc064925e43b8f87057ed83f14059535520d4d261d97f56c2525c51bf934a4d7"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:24:32.888605Z","signature_b64":"cuHYCvRJd4gNv3sVO4qsQaEF32hJse/ecU6GkuIiS3SyFtooeK5qIDa15dOdLYSmPk5q831oplJVNSiMTzEaBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"b8e7a386a9528f3e6a83548791850ed09769619238e0731158f5ba5b048d8368","last_reissued_at":"2026-07-05T09:24:32.888122Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:24:32.888122Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Interpreting Context Look-ups in Transformers: Investigating Attention-MLP Interactions","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Clement Neo, Fazl Barez, Shay B. Cohen","submitted_at":"2024-02-23T02:15:47Z","abstract_excerpt":"Understanding the inner workings of large language models (LLMs) is crucial for advancing their theoretical foundations and real-world applications. While the attention mechanism and multi-layer perceptrons (MLPs) have been studied independently, their interactions remain largely unexplored. This study investigates how attention heads and next-token neurons interact in LLMs to predict new words. We propose a methodology to identify next-token neurons, find prompts that highly activate them, and determine the upstream attention heads responsible. We then generate and evaluate explanations for t"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2402.15055","kind":"arxiv","version":2},"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/2402.15055/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":"2402.15055","created_at":"2026-07-05T09:24:32.888187+00:00"},{"alias_kind":"arxiv_version","alias_value":"2402.15055v2","created_at":"2026-07-05T09:24:32.888187+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.15055","created_at":"2026-07-05T09:24:32.888187+00:00"},{"alias_kind":"pith_short_12","alias_value":"XDT2HBVJKKHT","created_at":"2026-07-05T09:24:32.888187+00:00"},{"alias_kind":"pith_short_16","alias_value":"XDT2HBVJKKHT42UD","created_at":"2026-07-05T09:24:32.888187+00:00"},{"alias_kind":"pith_short_8","alias_value":"XDT2HBVJ","created_at":"2026-07-05T09:24:32.888187+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/XDT2HBVJKKHT42UDKSDZDBIO2C","json":"https://pith.science/pith/XDT2HBVJKKHT42UDKSDZDBIO2C.json","graph_json":"https://pith.science/api/pith-number/XDT2HBVJKKHT42UDKSDZDBIO2C/graph.json","events_json":"https://pith.science/api/pith-number/XDT2HBVJKKHT42UDKSDZDBIO2C/events.json","paper":"https://pith.science/paper/XDT2HBVJ"},"agent_actions":{"view_html":"https://pith.science/pith/XDT2HBVJKKHT42UDKSDZDBIO2C","download_json":"https://pith.science/pith/XDT2HBVJKKHT42UDKSDZDBIO2C.json","view_paper":"https://pith.science/paper/XDT2HBVJ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2402.15055&json=true","fetch_graph":"https://pith.science/api/pith-number/XDT2HBVJKKHT42UDKSDZDBIO2C/graph.json","fetch_events":"https://pith.science/api/pith-number/XDT2HBVJKKHT42UDKSDZDBIO2C/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/XDT2HBVJKKHT42UDKSDZDBIO2C/action/timestamp_anchor","attest_storage":"https://pith.science/pith/XDT2HBVJKKHT42UDKSDZDBIO2C/action/storage_attestation","attest_author":"https://pith.science/pith/XDT2HBVJKKHT42UDKSDZDBIO2C/action/author_attestation","sign_citation":"https://pith.science/pith/XDT2HBVJKKHT42UDKSDZDBIO2C/action/citation_signature","submit_replication":"https://pith.science/pith/XDT2HBVJKKHT42UDKSDZDBIO2C/action/replication_record"}},"created_at":"2026-07-05T09:24:32.888187+00:00","updated_at":"2026-07-05T09:24:32.888187+00:00"}