{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:DU4W7MVV5JLFP7YPGDLKIWS6UA","short_pith_number":"pith:DU4W7MVV","canonical_record":{"source":{"id":"2506.15553","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-06-18T15:26:43Z","cross_cats_sorted":[],"title_canon_sha256":"29b62ac19407831883ab6d46975ddfbed82559ca18d0581674407d9daf3d89d7","abstract_canon_sha256":"533b4dc040d573ae95b040541a10715dbb16a3b36aa235da8d2fa0602b18c1d1"},"schema_version":"1.0"},"canonical_sha256":"1d396fb2b5ea5657ff0f30d6a45a5ea01e472e98c51f38b55ba787c2f202c169","source":{"kind":"arxiv","id":"2506.15553","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2506.15553","created_at":"2026-07-05T11:23:45Z"},{"alias_kind":"arxiv_version","alias_value":"2506.15553v1","created_at":"2026-07-05T11:23:45Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.15553","created_at":"2026-07-05T11:23:45Z"},{"alias_kind":"pith_short_12","alias_value":"DU4W7MVV5JLF","created_at":"2026-07-05T11:23:45Z"},{"alias_kind":"pith_short_16","alias_value":"DU4W7MVV5JLFP7YP","created_at":"2026-07-05T11:23:45Z"},{"alias_kind":"pith_short_8","alias_value":"DU4W7MVV","created_at":"2026-07-05T11:23:45Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:DU4W7MVV5JLFP7YPGDLKIWS6UA","target":"record","payload":{"canonical_record":{"source":{"id":"2506.15553","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-06-18T15:26:43Z","cross_cats_sorted":[],"title_canon_sha256":"29b62ac19407831883ab6d46975ddfbed82559ca18d0581674407d9daf3d89d7","abstract_canon_sha256":"533b4dc040d573ae95b040541a10715dbb16a3b36aa235da8d2fa0602b18c1d1"},"schema_version":"1.0"},"canonical_sha256":"1d396fb2b5ea5657ff0f30d6a45a5ea01e472e98c51f38b55ba787c2f202c169","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:23:45.991803Z","signature_b64":"Lke/fE87VDBe5A/+GwH4Mvzp/6+vXYYudulrK6exMZyF7XvB8c/oKBxmAycAJ3Ha5pJBGtjtXpj8YDoy6maJAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"1d396fb2b5ea5657ff0f30d6a45a5ea01e472e98c51f38b55ba787c2f202c169","last_reissued_at":"2026-07-05T11:23:45.991297Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:23:45.991297Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2506.15553","source_version":1,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T11:23:45Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"4YVm2aLxSnb5zloZdUcEgdf0mHCiw3WPb+m4+MWQ8TJe5gvkrUq5J2LbO4V5uZGiiLusMVXLivjZtMviKmfGCw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-12T09:16:25.550618Z"},"content_sha256":"10cb8ebd086bfe726b24151c925c700659a66be6fc81ae4f584defba8818ecf5","schema_version":"1.0","event_id":"sha256:10cb8ebd086bfe726b24151c925c700659a66be6fc81ae4f584defba8818ecf5"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:DU4W7MVV5JLFP7YPGDLKIWS6UA","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Approximating Language Model Training Data from Weights","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Alexander M. Rush, John X. Morris, Junjie Oscar Yin, Vitaly Shmatikov, Woojeong Kim","submitted_at":"2025-06-18T15:26:43Z","abstract_excerpt":"Modern language models often have open weights but closed training data. We formalize the problem of data approximation from model weights and propose several baselines and metrics. We develop a gradient-based approach that selects the highest-matching data from a large public text corpus and show its effectiveness at recovering useful data given only weights of the original and finetuned models. Even when none of the true training data is known, our method is able to locate a small subset of public Web documents can be used to train a model to close to the original model performance given mod"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.15553","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/2506.15553/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":"2026-07-05T11:23:45Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"fZ/M/orRHhrj+LISOyzPRjCGG4KyEMpCUGgFhkREexvoQMZ+neESyEbxIDVYJyEhma2rUbYZR44DTMT6brvqBg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-12T09:16:25.551602Z"},"content_sha256":"8eac4393be7afd2cbef967f6db12c462352902d9c398d43d29d1293dbaa61fd8","schema_version":"1.0","event_id":"sha256:8eac4393be7afd2cbef967f6db12c462352902d9c398d43d29d1293dbaa61fd8"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/DU4W7MVV5JLFP7YPGDLKIWS6UA/bundle.json","state_url":"https://pith.science/pith/DU4W7MVV5JLFP7YPGDLKIWS6UA/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/DU4W7MVV5JLFP7YPGDLKIWS6UA/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":"2026-08-12T09:16:25Z","links":{"resolver":"https://pith.science/pith/DU4W7MVV5JLFP7YPGDLKIWS6UA","bundle":"https://pith.science/pith/DU4W7MVV5JLFP7YPGDLKIWS6UA/bundle.json","state":"https://pith.science/pith/DU4W7MVV5JLFP7YPGDLKIWS6UA/state.json","well_known_bundle":"https://pith.science/.well-known/pith/DU4W7MVV5JLFP7YPGDLKIWS6UA/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:DU4W7MVV5JLFP7YPGDLKIWS6UA","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":"533b4dc040d573ae95b040541a10715dbb16a3b36aa235da8d2fa0602b18c1d1","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-06-18T15:26:43Z","title_canon_sha256":"29b62ac19407831883ab6d46975ddfbed82559ca18d0581674407d9daf3d89d7"},"schema_version":"1.0","source":{"id":"2506.15553","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2506.15553","created_at":"2026-07-05T11:23:45Z"},{"alias_kind":"arxiv_version","alias_value":"2506.15553v1","created_at":"2026-07-05T11:23:45Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.15553","created_at":"2026-07-05T11:23:45Z"},{"alias_kind":"pith_short_12","alias_value":"DU4W7MVV5JLF","created_at":"2026-07-05T11:23:45Z"},{"alias_kind":"pith_short_16","alias_value":"DU4W7MVV5JLFP7YP","created_at":"2026-07-05T11:23:45Z"},{"alias_kind":"pith_short_8","alias_value":"DU4W7MVV","created_at":"2026-07-05T11:23:45Z"}],"graph_snapshots":[{"event_id":"sha256:8eac4393be7afd2cbef967f6db12c462352902d9c398d43d29d1293dbaa61fd8","target":"graph","created_at":"2026-07-05T11:23:45Z","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/2506.15553/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Modern language models often have open weights but closed training data. We formalize the problem of data approximation from model weights and propose several baselines and metrics. We develop a gradient-based approach that selects the highest-matching data from a large public text corpus and show its effectiveness at recovering useful data given only weights of the original and finetuned models. Even when none of the true training data is known, our method is able to locate a small subset of public Web documents can be used to train a model to close to the original model performance given mod","authors_text":"Alexander M. Rush, John X. Morris, Junjie Oscar Yin, Vitaly Shmatikov, Woojeong Kim","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-06-18T15:26:43Z","title":"Approximating Language Model Training Data from Weights"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.15553","kind":"arxiv","version":1},"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":"sha256:10cb8ebd086bfe726b24151c925c700659a66be6fc81ae4f584defba8818ecf5","target":"record","created_at":"2026-07-05T11:23:45Z","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":"533b4dc040d573ae95b040541a10715dbb16a3b36aa235da8d2fa0602b18c1d1","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-06-18T15:26:43Z","title_canon_sha256":"29b62ac19407831883ab6d46975ddfbed82559ca18d0581674407d9daf3d89d7"},"schema_version":"1.0","source":{"id":"2506.15553","kind":"arxiv","version":1}},"canonical_sha256":"1d396fb2b5ea5657ff0f30d6a45a5ea01e472e98c51f38b55ba787c2f202c169","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"1d396fb2b5ea5657ff0f30d6a45a5ea01e472e98c51f38b55ba787c2f202c169","first_computed_at":"2026-07-05T11:23:45.991297Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:23:45.991297Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"Lke/fE87VDBe5A/+GwH4Mvzp/6+vXYYudulrK6exMZyF7XvB8c/oKBxmAycAJ3Ha5pJBGtjtXpj8YDoy6maJAQ==","signature_status":"signed_v1","signed_at":"2026-07-05T11:23:45.991803Z","signed_message":"canonical_sha256_bytes"},"source_id":"2506.15553","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:10cb8ebd086bfe726b24151c925c700659a66be6fc81ae4f584defba8818ecf5","sha256:8eac4393be7afd2cbef967f6db12c462352902d9c398d43d29d1293dbaa61fd8"],"state_sha256":"a84b8e2b8d1cc5380dbd6be7000ed3cff1fce6a2942a8ef06bb0ee4e5578d963"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"vzyoqew6X2jT/Q2xjSS5l72YJvgcqL5peN4vMnIAs3EcGMHfAB/VsBaZfC5iZnyLKzjzRwtq8EHKHaWgyrbrAg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-12T09:16:25.557470Z","bundle_sha256":"808de0db19443df26acef00500ebed4385f90b37411f6b7509f3fb0e71b0235d"}}