{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:5X4576VLLETRGT47DZGQX3OGKF","short_pith_number":"pith:5X4576VL","schema_version":"1.0","canonical_sha256":"edf9dffaab5927134f9f1e4d0bedc65167bc521e942989795149ef07d820677e","source":{"kind":"arxiv","id":"2406.02585","version":1},"attestation_state":"computed","paper":{"title":"Contextual Counting: A Mechanistic Study of Transformers on a Quantitative Task","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","stat.ML"],"primary_cat":"cs.LG","authors_text":"Alberto Bietti, Bruno R\\'egaldo-Saint Blancard, Geraud Krawezik, Keiya Hirashima, Kyunghyun Cho, Liam Holden Parker, Mariel Pettee, Michael Eickenberg, Michael McCabe, Miles Cranmer, Nicholas Lourie, Ruben Ohana, Rudy Morel, Shirley Ho, Siavash Golkar","submitted_at":"2024-05-30T20:52:23Z","abstract_excerpt":"Transformers have revolutionized machine learning across diverse domains, yet understanding their behavior remains crucial, particularly in high-stakes applications. This paper introduces the contextual counting task, a novel toy problem aimed at enhancing our understanding of Transformers in quantitative and scientific contexts. This task requires precise localization and computation within datasets, akin to object detection or region-based scientific analysis. We present theoretical and empirical analysis using both causal and non-causal Transformer architectures, investigating the influence"},"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":"2406.02585","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-05-30T20:52:23Z","cross_cats_sorted":["cs.AI","stat.ML"],"title_canon_sha256":"4c3d3e3ffe7155ffd045439d560f692410c3eaaf46ca529de72f906c7eebc1ea","abstract_canon_sha256":"888546ae389831d7a3e79bd556e0fa7b60d707b3369583429a666687760989f5"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:27:27.346413Z","signature_b64":"5rjZX9IvnvFHFkoTl7VbvQGmhq/MybO+tqunmoztjBbJBG6U8U8KyE4t9eDc9QRUdBK/Q8ndE1UGq2bnX5H2Bw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"edf9dffaab5927134f9f1e4d0bedc65167bc521e942989795149ef07d820677e","last_reissued_at":"2026-07-05T08:27:27.345981Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:27:27.345981Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Contextual Counting: A Mechanistic Study of Transformers on a Quantitative Task","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","stat.ML"],"primary_cat":"cs.LG","authors_text":"Alberto Bietti, Bruno R\\'egaldo-Saint Blancard, Geraud Krawezik, Keiya Hirashima, Kyunghyun Cho, Liam Holden Parker, Mariel Pettee, Michael Eickenberg, Michael McCabe, Miles Cranmer, Nicholas Lourie, Ruben Ohana, Rudy Morel, Shirley Ho, Siavash Golkar","submitted_at":"2024-05-30T20:52:23Z","abstract_excerpt":"Transformers have revolutionized machine learning across diverse domains, yet understanding their behavior remains crucial, particularly in high-stakes applications. This paper introduces the contextual counting task, a novel toy problem aimed at enhancing our understanding of Transformers in quantitative and scientific contexts. This task requires precise localization and computation within datasets, akin to object detection or region-based scientific analysis. We present theoretical and empirical analysis using both causal and non-causal Transformer architectures, investigating the influence"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2406.02585","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/2406.02585/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":"2406.02585","created_at":"2026-07-05T08:27:27.346040+00:00"},{"alias_kind":"arxiv_version","alias_value":"2406.02585v1","created_at":"2026-07-05T08:27:27.346040+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2406.02585","created_at":"2026-07-05T08:27:27.346040+00:00"},{"alias_kind":"pith_short_12","alias_value":"5X4576VLLETR","created_at":"2026-07-05T08:27:27.346040+00:00"},{"alias_kind":"pith_short_16","alias_value":"5X4576VLLETRGT47","created_at":"2026-07-05T08:27:27.346040+00:00"},{"alias_kind":"pith_short_8","alias_value":"5X4576VL","created_at":"2026-07-05T08:27:27.346040+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2511.17699","citing_title":"Understanding Counting Mechanisms in Large Language and Vision-Language Models","ref_index":10,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/5X4576VLLETRGT47DZGQX3OGKF","json":"https://pith.science/pith/5X4576VLLETRGT47DZGQX3OGKF.json","graph_json":"https://pith.science/api/pith-number/5X4576VLLETRGT47DZGQX3OGKF/graph.json","events_json":"https://pith.science/api/pith-number/5X4576VLLETRGT47DZGQX3OGKF/events.json","paper":"https://pith.science/paper/5X4576VL"},"agent_actions":{"view_html":"https://pith.science/pith/5X4576VLLETRGT47DZGQX3OGKF","download_json":"https://pith.science/pith/5X4576VLLETRGT47DZGQX3OGKF.json","view_paper":"https://pith.science/paper/5X4576VL","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2406.02585&json=true","fetch_graph":"https://pith.science/api/pith-number/5X4576VLLETRGT47DZGQX3OGKF/graph.json","fetch_events":"https://pith.science/api/pith-number/5X4576VLLETRGT47DZGQX3OGKF/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/5X4576VLLETRGT47DZGQX3OGKF/action/timestamp_anchor","attest_storage":"https://pith.science/pith/5X4576VLLETRGT47DZGQX3OGKF/action/storage_attestation","attest_author":"https://pith.science/pith/5X4576VLLETRGT47DZGQX3OGKF/action/author_attestation","sign_citation":"https://pith.science/pith/5X4576VLLETRGT47DZGQX3OGKF/action/citation_signature","submit_replication":"https://pith.science/pith/5X4576VLLETRGT47DZGQX3OGKF/action/replication_record"}},"created_at":"2026-07-05T08:27:27.346040+00:00","updated_at":"2026-07-05T08:27:27.346040+00:00"}