{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:CIRQVQY5BQJ7VYTCLTLZYEQA7E","short_pith_number":"pith:CIRQVQY5","schema_version":"1.0","canonical_sha256":"12230ac31d0c13fae2625cd79c1200f90032586a827115d6fdeacfdda65aded7","source":{"kind":"arxiv","id":"2402.10891","version":1},"attestation_state":"computed","paper":{"title":"Instruction Diversity Drives Generalization To Unseen Tasks","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Dylan Zhang, Francois Charton, Justin Wang","submitted_at":"2024-02-16T18:47:21Z","abstract_excerpt":"Instruction tuning -- fine-tuning a large language model (LLM) on pairs of instructions and desired outcomes -- is an approach that enables pre-trained language models to perform real-world tasks and follow human instructions. Its practical success depends on the model learning a broader set of instructions than those it was trained on. Yet the factors that determine model generalization to such \\emph{unseen tasks} are not well understood. %To understand the driving factors of generalization, In this paper, we experiment with string rewrites, a symbolic task that serves as a building block for"},"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":"2402.10891","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2024-02-16T18:47:21Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"ebc5d9e50e70abe50e96466619c4d3f44b43439bdc78a3b4784885feb0eba18c","abstract_canon_sha256":"bcf0d1ff1618539629f9262af25eca458dc87a247cec55f926643b83e6a3da28"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:46:02.121424Z","signature_b64":"beGX93aveCcUmM3dU/z3MSDVwdlJ8Aha4ZhIOCqhToDawQjrsNDNODBj2sge+RgGwgd4Z2O57teLOJaDaAyHAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"12230ac31d0c13fae2625cd79c1200f90032586a827115d6fdeacfdda65aded7","last_reissued_at":"2026-07-05T07:46:02.121022Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:46:02.121022Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Instruction Diversity Drives Generalization To Unseen Tasks","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Dylan Zhang, Francois Charton, Justin Wang","submitted_at":"2024-02-16T18:47:21Z","abstract_excerpt":"Instruction tuning -- fine-tuning a large language model (LLM) on pairs of instructions and desired outcomes -- is an approach that enables pre-trained language models to perform real-world tasks and follow human instructions. Its practical success depends on the model learning a broader set of instructions than those it was trained on. Yet the factors that determine model generalization to such \\emph{unseen tasks} are not well understood. %To understand the driving factors of generalization, In this paper, we experiment with string rewrites, a symbolic task that serves as a building block for"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2402.10891","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/2402.10891/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":"2402.10891","created_at":"2026-07-05T07:46:02.121079+00:00"},{"alias_kind":"arxiv_version","alias_value":"2402.10891v1","created_at":"2026-07-05T07:46:02.121079+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.10891","created_at":"2026-07-05T07:46:02.121079+00:00"},{"alias_kind":"pith_short_12","alias_value":"CIRQVQY5BQJ7","created_at":"2026-07-05T07:46:02.121079+00:00"},{"alias_kind":"pith_short_16","alias_value":"CIRQVQY5BQJ7VYTC","created_at":"2026-07-05T07:46:02.121079+00:00"},{"alias_kind":"pith_short_8","alias_value":"CIRQVQY5","created_at":"2026-07-05T07:46:02.121079+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2507.11687","citing_title":"MetaLint: Easy-to-Hard Generalization for Code Linting","ref_index":5,"is_internal_anchor":false},{"citing_arxiv_id":"2605.05115","citing_title":"Manifold Steering Reveals the Shared Geometry of Neural Network Representation and Behavior","ref_index":275,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/CIRQVQY5BQJ7VYTCLTLZYEQA7E","json":"https://pith.science/pith/CIRQVQY5BQJ7VYTCLTLZYEQA7E.json","graph_json":"https://pith.science/api/pith-number/CIRQVQY5BQJ7VYTCLTLZYEQA7E/graph.json","events_json":"https://pith.science/api/pith-number/CIRQVQY5BQJ7VYTCLTLZYEQA7E/events.json","paper":"https://pith.science/paper/CIRQVQY5"},"agent_actions":{"view_html":"https://pith.science/pith/CIRQVQY5BQJ7VYTCLTLZYEQA7E","download_json":"https://pith.science/pith/CIRQVQY5BQJ7VYTCLTLZYEQA7E.json","view_paper":"https://pith.science/paper/CIRQVQY5","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2402.10891&json=true","fetch_graph":"https://pith.science/api/pith-number/CIRQVQY5BQJ7VYTCLTLZYEQA7E/graph.json","fetch_events":"https://pith.science/api/pith-number/CIRQVQY5BQJ7VYTCLTLZYEQA7E/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/CIRQVQY5BQJ7VYTCLTLZYEQA7E/action/timestamp_anchor","attest_storage":"https://pith.science/pith/CIRQVQY5BQJ7VYTCLTLZYEQA7E/action/storage_attestation","attest_author":"https://pith.science/pith/CIRQVQY5BQJ7VYTCLTLZYEQA7E/action/author_attestation","sign_citation":"https://pith.science/pith/CIRQVQY5BQJ7VYTCLTLZYEQA7E/action/citation_signature","submit_replication":"https://pith.science/pith/CIRQVQY5BQJ7VYTCLTLZYEQA7E/action/replication_record"}},"created_at":"2026-07-05T07:46:02.121079+00:00","updated_at":"2026-07-05T07:46:02.121079+00:00"}