{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:QRSBL2JQJUEZHVGM67JJL4GUGJ","short_pith_number":"pith:QRSBL2JQ","schema_version":"1.0","canonical_sha256":"846415e9304d0993d4ccf7d295f0d43257d3f960e31ea70993b996351b0c74e3","source":{"kind":"arxiv","id":"2404.02418","version":2},"attestation_state":"computed","paper":{"title":"Auxiliary task demands mask the capabilities of smaller language models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Jennifer Hu, Michael C. Frank","submitted_at":"2024-04-03T02:56:52Z","abstract_excerpt":"Developmental psychologists have argued about when cognitive capacities such as language understanding or theory of mind emerge. These debates often hinge on the concept of \"task demands\" -- the auxiliary challenges associated with performing a particular evaluation -- that may mask the child's underlying ability. The same issues arise when measuring the capacities of language models (LMs): performance on a task is a function of the model's underlying knowledge, combined with the model's ability to interpret and perform the task given its available resources. Here, we show that for analogical "},"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":"2404.02418","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-04-03T02:56:52Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"4817ef2a72566972d029da0d8d9dd0a7e811efd23bf624dd4cf6b5346acbeaee","abstract_canon_sha256":"b6d82239daceaad14d0ad6d71a0db9239e152aefda535b6e00eb6cba3576c4e8"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:49:52.948065Z","signature_b64":"RFffFXj5QtGkdYqDBwL8d90sOWOal/iFZgrRdoyvyTReBxkvMPxGT9evJjJXc60DpoUAeIxn6Q/M1i74PLsLAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"846415e9304d0993d4ccf7d295f0d43257d3f960e31ea70993b996351b0c74e3","last_reissued_at":"2026-07-05T08:49:52.947645Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:49:52.947645Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Auxiliary task demands mask the capabilities of smaller language models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Jennifer Hu, Michael C. Frank","submitted_at":"2024-04-03T02:56:52Z","abstract_excerpt":"Developmental psychologists have argued about when cognitive capacities such as language understanding or theory of mind emerge. These debates often hinge on the concept of \"task demands\" -- the auxiliary challenges associated with performing a particular evaluation -- that may mask the child's underlying ability. The same issues arise when measuring the capacities of language models (LMs): performance on a task is a function of the model's underlying knowledge, combined with the model's ability to interpret and perform the task given its available resources. Here, we show that for analogical "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2404.02418","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/2404.02418/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":"2404.02418","created_at":"2026-07-05T08:49:52.947708+00:00"},{"alias_kind":"arxiv_version","alias_value":"2404.02418v2","created_at":"2026-07-05T08:49:52.947708+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2404.02418","created_at":"2026-07-05T08:49:52.947708+00:00"},{"alias_kind":"pith_short_12","alias_value":"QRSBL2JQJUEZ","created_at":"2026-07-05T08:49:52.947708+00:00"},{"alias_kind":"pith_short_16","alias_value":"QRSBL2JQJUEZHVGM","created_at":"2026-07-05T08:49:52.947708+00:00"},{"alias_kind":"pith_short_8","alias_value":"QRSBL2JQ","created_at":"2026-07-05T08:49:52.947708+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":6,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.18284","citing_title":"Breaking the Solver Bottleneck: Training Task Generators at the Learnable Frontier","ref_index":117,"is_internal_anchor":false},{"citing_arxiv_id":"2606.05497","citing_title":"LEVANTE-bench: Multi-Scale Comparison of VLMs to Children Using Cognitive Tasks (or, \"Is Your VLM Smarter Than a 5th Grader?\")","ref_index":57,"is_internal_anchor":false},{"citing_arxiv_id":"2606.31916","citing_title":"Theory of Mind and Persuasion Beyond Conversation: Assessing the Capacity of LLMs to Induce Belief States via Planning and Action","ref_index":89,"is_internal_anchor":false},{"citing_arxiv_id":"2605.21683","citing_title":"Investigating Concept Alignment Using Implausible Category Members","ref_index":14,"is_internal_anchor":false},{"citing_arxiv_id":"2602.01997","citing_title":"On the Limits of Layer Pruning for Generative Reasoning in Large Language Models","ref_index":14,"is_internal_anchor":false},{"citing_arxiv_id":"2605.09042","citing_title":"Evaluating Pragmatic Reasoning in Large Language Models: Evidence from Scalar Diversity","ref_index":35,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/QRSBL2JQJUEZHVGM67JJL4GUGJ","json":"https://pith.science/pith/QRSBL2JQJUEZHVGM67JJL4GUGJ.json","graph_json":"https://pith.science/api/pith-number/QRSBL2JQJUEZHVGM67JJL4GUGJ/graph.json","events_json":"https://pith.science/api/pith-number/QRSBL2JQJUEZHVGM67JJL4GUGJ/events.json","paper":"https://pith.science/paper/QRSBL2JQ"},"agent_actions":{"view_html":"https://pith.science/pith/QRSBL2JQJUEZHVGM67JJL4GUGJ","download_json":"https://pith.science/pith/QRSBL2JQJUEZHVGM67JJL4GUGJ.json","view_paper":"https://pith.science/paper/QRSBL2JQ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2404.02418&json=true","fetch_graph":"https://pith.science/api/pith-number/QRSBL2JQJUEZHVGM67JJL4GUGJ/graph.json","fetch_events":"https://pith.science/api/pith-number/QRSBL2JQJUEZHVGM67JJL4GUGJ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/QRSBL2JQJUEZHVGM67JJL4GUGJ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/QRSBL2JQJUEZHVGM67JJL4GUGJ/action/storage_attestation","attest_author":"https://pith.science/pith/QRSBL2JQJUEZHVGM67JJL4GUGJ/action/author_attestation","sign_citation":"https://pith.science/pith/QRSBL2JQJUEZHVGM67JJL4GUGJ/action/citation_signature","submit_replication":"https://pith.science/pith/QRSBL2JQJUEZHVGM67JJL4GUGJ/action/replication_record"}},"created_at":"2026-07-05T08:49:52.947708+00:00","updated_at":"2026-07-05T08:49:52.947708+00:00"}