{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:LLNNHA5ZLJ2WRMKVP5NM6537V3","short_pith_number":"pith:LLNNHA5Z","canonical_record":{"source":{"id":"2605.30537","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-28T20:12:45Z","cross_cats_sorted":[],"title_canon_sha256":"bfaf1fbb882daa65d7db49e656c62e2d2155795cae9db6370646cfe30ef985d2","abstract_canon_sha256":"94cd3262335a86cda3f4be9e6fee1f85992cabcafeafac47bef3348d29be6269"},"schema_version":"1.0"},"canonical_sha256":"5adad383b95a7568b1557f5acf777faee37c89322f97d4153874cb65d5153a28","source":{"kind":"arxiv","id":"2605.30537","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.30537","created_at":"2026-06-01T01:02:59Z"},{"alias_kind":"arxiv_version","alias_value":"2605.30537v1","created_at":"2026-06-01T01:02:59Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.30537","created_at":"2026-06-01T01:02:59Z"},{"alias_kind":"pith_short_12","alias_value":"LLNNHA5ZLJ2W","created_at":"2026-06-01T01:02:59Z"},{"alias_kind":"pith_short_16","alias_value":"LLNNHA5ZLJ2WRMKV","created_at":"2026-06-01T01:02:59Z"},{"alias_kind":"pith_short_8","alias_value":"LLNNHA5Z","created_at":"2026-06-01T01:02:59Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:LLNNHA5ZLJ2WRMKVP5NM6537V3","target":"record","payload":{"canonical_record":{"source":{"id":"2605.30537","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-28T20:12:45Z","cross_cats_sorted":[],"title_canon_sha256":"bfaf1fbb882daa65d7db49e656c62e2d2155795cae9db6370646cfe30ef985d2","abstract_canon_sha256":"94cd3262335a86cda3f4be9e6fee1f85992cabcafeafac47bef3348d29be6269"},"schema_version":"1.0"},"canonical_sha256":"5adad383b95a7568b1557f5acf777faee37c89322f97d4153874cb65d5153a28","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-01T01:02:59.663634Z","signature_b64":"nK4Ud7B583WYXcSyFvMwtn5IfT3ad+ijdAkbmRwKWUjwZTMXeMpK9sp0cvlIQlAtecaxypZzlVDuWnIabb9RBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"5adad383b95a7568b1557f5acf777faee37c89322f97d4153874cb65d5153a28","last_reissued_at":"2026-06-01T01:02:59.662742Z","signature_status":"signed_v1","first_computed_at":"2026-06-01T01:02:59.662742Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2605.30537","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":"2026-06-01T01:02:59Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"xSCVdKMuB4XVSZAyH+eIxk73XtAcyvihi6etzkQQeTXw2M3ZU1KLqA0h1sJnOUXXUKdVOm1xhQ1D1yT+oZ0HDQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-29T22:31:56.800035Z"},"content_sha256":"f14900ec99cea87757cf4c312293fcf1955f21c855c944ff7c71cfc615517029","schema_version":"1.0","event_id":"sha256:f14900ec99cea87757cf4c312293fcf1955f21c855c944ff7c71cfc615517029"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:LLNNHA5ZLJ2WRMKVP5NM6537V3","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"The Long-Term Effects of Data Selection in LLM Fine-Tuning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Aoxiong Zeng, Xiangquan Yang, Yuxin Yang","submitted_at":"2026-05-28T20:12:45Z","abstract_excerpt":"Data selection is increasingly used to reduce the cost of large language model (LLM) fine-tuning, with recent methods prioritizing samples by current utility, diversity, quality, or influence. This paper studies a different question: when fine-tuning occurs over multiple stages, can selection strategies that look optimal now make the model less adaptable later? We introduce a long-horizon view of LLM data selection in which a selector is evaluated not only by immediate task performance, but also by future adaptation speed, forgetting, capability imbalance, and out-of-distribution robustness. W"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.30537","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/2605.30537/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":"2026-06-01T01:02:59Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"9QDrrBNpECDXJtdFB0GGV19fZFxzTxQXT4FGrdVycAtwEUaXSDCcy02wS5ttk1hmjn+8j6JXgpnAP51xBhTICw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-29T22:31:56.800419Z"},"content_sha256":"914f64792981b38ce54b33998e3d54b804e9eb67b9cdede5e74e21879135d64e","schema_version":"1.0","event_id":"sha256:914f64792981b38ce54b33998e3d54b804e9eb67b9cdede5e74e21879135d64e"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/LLNNHA5ZLJ2WRMKVP5NM6537V3/bundle.json","state_url":"https://pith.science/pith/LLNNHA5ZLJ2WRMKVP5NM6537V3/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/LLNNHA5ZLJ2WRMKVP5NM6537V3/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":"2026-06-29T22:31:56Z","links":{"resolver":"https://pith.science/pith/LLNNHA5ZLJ2WRMKVP5NM6537V3","bundle":"https://pith.science/pith/LLNNHA5ZLJ2WRMKVP5NM6537V3/bundle.json","state":"https://pith.science/pith/LLNNHA5ZLJ2WRMKVP5NM6537V3/state.json","well_known_bundle":"https://pith.science/.well-known/pith/LLNNHA5ZLJ2WRMKVP5NM6537V3/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:LLNNHA5ZLJ2WRMKVP5NM6537V3","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":"94cd3262335a86cda3f4be9e6fee1f85992cabcafeafac47bef3348d29be6269","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-28T20:12:45Z","title_canon_sha256":"bfaf1fbb882daa65d7db49e656c62e2d2155795cae9db6370646cfe30ef985d2"},"schema_version":"1.0","source":{"id":"2605.30537","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.30537","created_at":"2026-06-01T01:02:59Z"},{"alias_kind":"arxiv_version","alias_value":"2605.30537v1","created_at":"2026-06-01T01:02:59Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.30537","created_at":"2026-06-01T01:02:59Z"},{"alias_kind":"pith_short_12","alias_value":"LLNNHA5ZLJ2W","created_at":"2026-06-01T01:02:59Z"},{"alias_kind":"pith_short_16","alias_value":"LLNNHA5ZLJ2WRMKV","created_at":"2026-06-01T01:02:59Z"},{"alias_kind":"pith_short_8","alias_value":"LLNNHA5Z","created_at":"2026-06-01T01:02:59Z"}],"graph_snapshots":[{"event_id":"sha256:914f64792981b38ce54b33998e3d54b804e9eb67b9cdede5e74e21879135d64e","target":"graph","created_at":"2026-06-01T01:02:59Z","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/2605.30537/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Data selection is increasingly used to reduce the cost of large language model (LLM) fine-tuning, with recent methods prioritizing samples by current utility, diversity, quality, or influence. This paper studies a different question: when fine-tuning occurs over multiple stages, can selection strategies that look optimal now make the model less adaptable later? We introduce a long-horizon view of LLM data selection in which a selector is evaluated not only by immediate task performance, but also by future adaptation speed, forgetting, capability imbalance, and out-of-distribution robustness. W","authors_text":"Aoxiong Zeng, Xiangquan Yang, Yuxin Yang","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-28T20:12:45Z","title":"The Long-Term Effects of Data Selection in LLM Fine-Tuning"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.30537","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:f14900ec99cea87757cf4c312293fcf1955f21c855c944ff7c71cfc615517029","target":"record","created_at":"2026-06-01T01:02:59Z","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":"94cd3262335a86cda3f4be9e6fee1f85992cabcafeafac47bef3348d29be6269","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-28T20:12:45Z","title_canon_sha256":"bfaf1fbb882daa65d7db49e656c62e2d2155795cae9db6370646cfe30ef985d2"},"schema_version":"1.0","source":{"id":"2605.30537","kind":"arxiv","version":1}},"canonical_sha256":"5adad383b95a7568b1557f5acf777faee37c89322f97d4153874cb65d5153a28","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"5adad383b95a7568b1557f5acf777faee37c89322f97d4153874cb65d5153a28","first_computed_at":"2026-06-01T01:02:59.662742Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-06-01T01:02:59.662742Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"nK4Ud7B583WYXcSyFvMwtn5IfT3ad+ijdAkbmRwKWUjwZTMXeMpK9sp0cvlIQlAtecaxypZzlVDuWnIabb9RBw==","signature_status":"signed_v1","signed_at":"2026-06-01T01:02:59.663634Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.30537","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:f14900ec99cea87757cf4c312293fcf1955f21c855c944ff7c71cfc615517029","sha256:914f64792981b38ce54b33998e3d54b804e9eb67b9cdede5e74e21879135d64e"],"state_sha256":"2e692215927b8a2e459891681119893dd5db92f515c9fba9024bebfd9228a188"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"DfCLKOw3vQ0g96tMgP1McN3HcCTgc+rLVy4PrHNmKyXwFI3AqXkyCcaGe4tYijfsORW3wACaw9xUCs4EsBrzCQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-06-29T22:31:56.802789Z","bundle_sha256":"747ae61bf611eff5712c845ac927b2fcf3d489915899dcdd13d4f27ff5fcd064"}}