{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:BATXQZ7XQFGLFQ66PE5M52DCUD","short_pith_number":"pith:BATXQZ7X","canonical_record":{"source":{"id":"2603.04828","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2026-03-05T05:21:51Z","cross_cats_sorted":[],"title_canon_sha256":"28a6c8382b2eae2d58042c64a6a0fe35cb25751f3e6aa9ff250eff15f39d5b04","abstract_canon_sha256":"ad7e636f9c1ab5cd6b4e55a1a43cbfe81b8caf7115979b4070a8e897b23fd27e"},"schema_version":"1.0"},"canonical_sha256":"08277867f7814cb2c3de793acee862a0debc79a836179a59a46d6a84bf78d732","source":{"kind":"arxiv","id":"2603.04828","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2603.04828","created_at":"2026-06-02T01:03:44Z"},{"alias_kind":"arxiv_version","alias_value":"2603.04828v2","created_at":"2026-06-02T01:03:44Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2603.04828","created_at":"2026-06-02T01:03:44Z"},{"alias_kind":"pith_short_12","alias_value":"BATXQZ7XQFGL","created_at":"2026-06-02T01:03:44Z"},{"alias_kind":"pith_short_16","alias_value":"BATXQZ7XQFGLFQ66","created_at":"2026-06-02T01:03:44Z"},{"alias_kind":"pith_short_8","alias_value":"BATXQZ7X","created_at":"2026-06-02T01:03:44Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:BATXQZ7XQFGLFQ66PE5M52DCUD","target":"record","payload":{"canonical_record":{"source":{"id":"2603.04828","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2026-03-05T05:21:51Z","cross_cats_sorted":[],"title_canon_sha256":"28a6c8382b2eae2d58042c64a6a0fe35cb25751f3e6aa9ff250eff15f39d5b04","abstract_canon_sha256":"ad7e636f9c1ab5cd6b4e55a1a43cbfe81b8caf7115979b4070a8e897b23fd27e"},"schema_version":"1.0"},"canonical_sha256":"08277867f7814cb2c3de793acee862a0debc79a836179a59a46d6a84bf78d732","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-02T01:03:44.628180Z","signature_b64":"Ssw0B60Ae9gbA9EexlEXWZ01QqKhZDCtrodXKjW6Ivr64Ky6nukul6Jb3mIRfx3d5sqNY2K8dSnHkwzW0gH+BQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"08277867f7814cb2c3de793acee862a0debc79a836179a59a46d6a84bf78d732","last_reissued_at":"2026-06-02T01:03:44.627665Z","signature_status":"signed_v1","first_computed_at":"2026-06-02T01:03:44.627665Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2603.04828","source_version":2,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-06-02T01:03:44Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"RviUD7s6MV/JPhNsP8cd8Lyocy6NTsbJNfYChtPsIiGBOMbQGbJ5K2n1CCGZ+LruVYwooKnvVRLmh7EVa5XkBg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-03T05:36:31.151943Z"},"content_sha256":"1ffb950677e27cd1c438abbe499ec8cb5bd3965f41904035ee5f4bfbc8eda48b","schema_version":"1.0","event_id":"sha256:1ffb950677e27cd1c438abbe499ec8cb5bd3965f41904035ee5f4bfbc8eda48b"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:BATXQZ7XQFGLFQ66PE5M52DCUD","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"From Unfamiliar to Familiar: Detecting Pre-training Data via Gradient Deviations in Large Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Hainan Zhang, Lingxiang Wang, Ruiqi Zhang, Yanyan Lan, Zhiming Zheng","submitted_at":"2026-03-05T05:21:51Z","abstract_excerpt":"Pre-training data detection for LLMs is essential for addressing copyright concerns and mitigating benchmark contamination. Existing methods mainly focus on the likelihood-based statistical features or heuristic signals before and after fine-tuning, but the former are susceptible to word frequency bias in corpora, and the latter strongly depend on the similarity of fine-tuning data. From an optimization perspective, we observe that during training, samples transition from unfamiliar to familiar in a manner reflected by systematic differences in gradient behavior. Familiar samples exhibit small"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2603.04828","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/2603.04828/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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-06-02T01:03:44Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"icyYi6r/epX+Lr+PQYoBdl1UUsC4tcp+4KutUHvp7/5Ov5hj63XjncgJt6i3VqyEgsX2bdlkN3WeJMPgucMUBA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-03T05:36:31.152338Z"},"content_sha256":"077f4cc4ab7b455aacf1e2a1b54144513f8ace920e75ddba49152128013b1be5","schema_version":"1.0","event_id":"sha256:077f4cc4ab7b455aacf1e2a1b54144513f8ace920e75ddba49152128013b1be5"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/BATXQZ7XQFGLFQ66PE5M52DCUD/bundle.json","state_url":"https://pith.science/pith/BATXQZ7XQFGLFQ66PE5M52DCUD/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/BATXQZ7XQFGLFQ66PE5M52DCUD/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-06-03T05:36:31Z","links":{"resolver":"https://pith.science/pith/BATXQZ7XQFGLFQ66PE5M52DCUD","bundle":"https://pith.science/pith/BATXQZ7XQFGLFQ66PE5M52DCUD/bundle.json","state":"https://pith.science/pith/BATXQZ7XQFGLFQ66PE5M52DCUD/state.json","well_known_bundle":"https://pith.science/.well-known/pith/BATXQZ7XQFGLFQ66PE5M52DCUD/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:BATXQZ7XQFGLFQ66PE5M52DCUD","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"ad7e636f9c1ab5cd6b4e55a1a43cbfe81b8caf7115979b4070a8e897b23fd27e","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2026-03-05T05:21:51Z","title_canon_sha256":"28a6c8382b2eae2d58042c64a6a0fe35cb25751f3e6aa9ff250eff15f39d5b04"},"schema_version":"1.0","source":{"id":"2603.04828","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2603.04828","created_at":"2026-06-02T01:03:44Z"},{"alias_kind":"arxiv_version","alias_value":"2603.04828v2","created_at":"2026-06-02T01:03:44Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2603.04828","created_at":"2026-06-02T01:03:44Z"},{"alias_kind":"pith_short_12","alias_value":"BATXQZ7XQFGL","created_at":"2026-06-02T01:03:44Z"},{"alias_kind":"pith_short_16","alias_value":"BATXQZ7XQFGLFQ66","created_at":"2026-06-02T01:03:44Z"},{"alias_kind":"pith_short_8","alias_value":"BATXQZ7X","created_at":"2026-06-02T01:03:44Z"}],"graph_snapshots":[{"event_id":"sha256:077f4cc4ab7b455aacf1e2a1b54144513f8ace920e75ddba49152128013b1be5","target":"graph","created_at":"2026-06-02T01:03:44Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2603.04828/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Pre-training data detection for LLMs is essential for addressing copyright concerns and mitigating benchmark contamination. Existing methods mainly focus on the likelihood-based statistical features or heuristic signals before and after fine-tuning, but the former are susceptible to word frequency bias in corpora, and the latter strongly depend on the similarity of fine-tuning data. From an optimization perspective, we observe that during training, samples transition from unfamiliar to familiar in a manner reflected by systematic differences in gradient behavior. Familiar samples exhibit small","authors_text":"Hainan Zhang, Lingxiang Wang, Ruiqi Zhang, Yanyan Lan, Zhiming Zheng","cross_cats":[],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2026-03-05T05:21:51Z","title":"From Unfamiliar to Familiar: Detecting Pre-training Data via Gradient Deviations in Large Language Models"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2603.04828","kind":"arxiv","version":2},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:1ffb950677e27cd1c438abbe499ec8cb5bd3965f41904035ee5f4bfbc8eda48b","target":"record","created_at":"2026-06-02T01:03:44Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"ad7e636f9c1ab5cd6b4e55a1a43cbfe81b8caf7115979b4070a8e897b23fd27e","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2026-03-05T05:21:51Z","title_canon_sha256":"28a6c8382b2eae2d58042c64a6a0fe35cb25751f3e6aa9ff250eff15f39d5b04"},"schema_version":"1.0","source":{"id":"2603.04828","kind":"arxiv","version":2}},"canonical_sha256":"08277867f7814cb2c3de793acee862a0debc79a836179a59a46d6a84bf78d732","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"08277867f7814cb2c3de793acee862a0debc79a836179a59a46d6a84bf78d732","first_computed_at":"2026-06-02T01:03:44.627665Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-06-02T01:03:44.627665Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"Ssw0B60Ae9gbA9EexlEXWZ01QqKhZDCtrodXKjW6Ivr64Ky6nukul6Jb3mIRfx3d5sqNY2K8dSnHkwzW0gH+BQ==","signature_status":"signed_v1","signed_at":"2026-06-02T01:03:44.628180Z","signed_message":"canonical_sha256_bytes"},"source_id":"2603.04828","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:1ffb950677e27cd1c438abbe499ec8cb5bd3965f41904035ee5f4bfbc8eda48b","sha256:077f4cc4ab7b455aacf1e2a1b54144513f8ace920e75ddba49152128013b1be5"],"state_sha256":"f704b67f90ce1f859962b05441777ed8276898f0b9b2e865b6d117295951615b"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"h0c+VxaiCv4vJ7BOffOPpg7GgYeBbrhJ34D9SBv2/8fslCWeLwoH1C9Sn34pZb0mHfYlHnig5E5+6Z79Pdk4Dg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-06-03T05:36:31.154345Z","bundle_sha256":"dc60f88aa54c38c2df48b724deff1550a8e12cead5f2cb266cb60825252e5a37"}}