{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:O7IGY3QH6YTQSTIXSDGBUQITOZ","short_pith_number":"pith:O7IGY3QH","schema_version":"1.0","canonical_sha256":"77d06c6e07f627094d1790cc1a411376474baaebbcd44ddfcd000a0c207e25d8","source":{"kind":"arxiv","id":"2607.09692","version":1},"attestation_state":"computed","paper":{"title":"Reference-Based Distillation Detection in LLMs","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.LG","authors_text":"Akshay Anand, Bob Rotsted, Michael Duan, Rajat Rawat, Sewon Min, Sizhe Chen","submitted_at":"2026-06-19T06:59:19Z","abstract_excerpt":"Model distillation -- training on outputs from stronger third-party models -- is widely used to boost performance, but raises concerns about unfair advantages and policy violations. This motivates a fundamental question: can we detect whether a model was distilled from another? We show that, while identifying a teacher model from a student in isolation is highly challenging, it becomes tractable in a reference-based setting: given a model and an earlier-generation checkpoint from the same lineage, we can identify the teacher model used to train the later checkpoint. We introduce a distillation"},"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":"2607.09692","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-06-19T06:59:19Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"30aee3413161a73a0d27e7884fdfaaef397a13432cecd42b324596e76cd15faf","abstract_canon_sha256":"0a71bd16aab6208da5b42c5754d2e3184ae32cac6159a5e14fe7573d5f89f7fe"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-14T00:18:30.743468Z","signature_b64":"E6fwn4es0HG/AmzYulV4tKnOSQz/j0XEZAetiueRVuNumXyv4+bVcVJUdpAFhZ9Wy6ql/Th3LjS8GmP7Fqc/DA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"77d06c6e07f627094d1790cc1a411376474baaebbcd44ddfcd000a0c207e25d8","last_reissued_at":"2026-07-14T00:18:30.742552Z","signature_status":"signed_v1","first_computed_at":"2026-07-14T00:18:30.742552Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Reference-Based Distillation Detection in LLMs","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.LG","authors_text":"Akshay Anand, Bob Rotsted, Michael Duan, Rajat Rawat, Sewon Min, Sizhe Chen","submitted_at":"2026-06-19T06:59:19Z","abstract_excerpt":"Model distillation -- training on outputs from stronger third-party models -- is widely used to boost performance, but raises concerns about unfair advantages and policy violations. This motivates a fundamental question: can we detect whether a model was distilled from another? We show that, while identifying a teacher model from a student in isolation is highly challenging, it becomes tractable in a reference-based setting: given a model and an earlier-generation checkpoint from the same lineage, we can identify the teacher model used to train the later checkpoint. We introduce a distillation"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.09692","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/2607.09692/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":"2607.09692","created_at":"2026-07-14T00:18:30.743009+00:00"},{"alias_kind":"arxiv_version","alias_value":"2607.09692v1","created_at":"2026-07-14T00:18:30.743009+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.09692","created_at":"2026-07-14T00:18:30.743009+00:00"},{"alias_kind":"pith_short_12","alias_value":"O7IGY3QH6YTQ","created_at":"2026-07-14T00:18:30.743009+00:00"},{"alias_kind":"pith_short_16","alias_value":"O7IGY3QH6YTQSTIX","created_at":"2026-07-14T00:18:30.743009+00:00"},{"alias_kind":"pith_short_8","alias_value":"O7IGY3QH","created_at":"2026-07-14T00:18:30.743009+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/O7IGY3QH6YTQSTIXSDGBUQITOZ","json":"https://pith.science/pith/O7IGY3QH6YTQSTIXSDGBUQITOZ.json","graph_json":"https://pith.science/api/pith-number/O7IGY3QH6YTQSTIXSDGBUQITOZ/graph.json","events_json":"https://pith.science/api/pith-number/O7IGY3QH6YTQSTIXSDGBUQITOZ/events.json","paper":"https://pith.science/paper/O7IGY3QH"},"agent_actions":{"view_html":"https://pith.science/pith/O7IGY3QH6YTQSTIXSDGBUQITOZ","download_json":"https://pith.science/pith/O7IGY3QH6YTQSTIXSDGBUQITOZ.json","view_paper":"https://pith.science/paper/O7IGY3QH","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2607.09692&json=true","fetch_graph":"https://pith.science/api/pith-number/O7IGY3QH6YTQSTIXSDGBUQITOZ/graph.json","fetch_events":"https://pith.science/api/pith-number/O7IGY3QH6YTQSTIXSDGBUQITOZ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/O7IGY3QH6YTQSTIXSDGBUQITOZ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/O7IGY3QH6YTQSTIXSDGBUQITOZ/action/storage_attestation","attest_author":"https://pith.science/pith/O7IGY3QH6YTQSTIXSDGBUQITOZ/action/author_attestation","sign_citation":"https://pith.science/pith/O7IGY3QH6YTQSTIXSDGBUQITOZ/action/citation_signature","submit_replication":"https://pith.science/pith/O7IGY3QH6YTQSTIXSDGBUQITOZ/action/replication_record"}},"created_at":"2026-07-14T00:18:30.743009+00:00","updated_at":"2026-07-14T00:18:30.743009+00:00"}