{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:S2HT6JMPE6EJFHYIOC5MLY757P","short_pith_number":"pith:S2HT6JMP","canonical_record":{"source":{"id":"2505.19893","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-05-26T12:23:26Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"283e93c2b269f37f0d34aeb04f1754f7f049de5899bc5675c43fa3f07a148691","abstract_canon_sha256":"84443c770689619c652869503566e6a8516d78ff83389a010597ac711381aab4"},"schema_version":"1.0"},"canonical_sha256":"968f3f258f2788929f0870bac5e3fdfbe1cc74171d0c94ec105959d507e2acc7","source":{"kind":"arxiv","id":"2505.19893","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2505.19893","created_at":"2026-07-05T11:09:49Z"},{"alias_kind":"arxiv_version","alias_value":"2505.19893v1","created_at":"2026-07-05T11:09:49Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.19893","created_at":"2026-07-05T11:09:49Z"},{"alias_kind":"pith_short_12","alias_value":"S2HT6JMPE6EJ","created_at":"2026-07-05T11:09:49Z"},{"alias_kind":"pith_short_16","alias_value":"S2HT6JMPE6EJFHYI","created_at":"2026-07-05T11:09:49Z"},{"alias_kind":"pith_short_8","alias_value":"S2HT6JMP","created_at":"2026-07-05T11:09:49Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:S2HT6JMPE6EJFHYIOC5MLY757P","target":"record","payload":{"canonical_record":{"source":{"id":"2505.19893","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-05-26T12:23:26Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"283e93c2b269f37f0d34aeb04f1754f7f049de5899bc5675c43fa3f07a148691","abstract_canon_sha256":"84443c770689619c652869503566e6a8516d78ff83389a010597ac711381aab4"},"schema_version":"1.0"},"canonical_sha256":"968f3f258f2788929f0870bac5e3fdfbe1cc74171d0c94ec105959d507e2acc7","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:09:49.569534Z","signature_b64":"YPBEGj9HVuwyLDcDJEAgR3ZUIq6WDIJ6IxhxbD7aifjQATlLWQkTY1P+BagZBpQkJF0GTT4oJ7PGxFs2Iru3Aw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"968f3f258f2788929f0870bac5e3fdfbe1cc74171d0c94ec105959d507e2acc7","last_reissued_at":"2026-07-05T11:09:49.569024Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:09:49.569024Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2505.19893","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":"2026-07-05T11:09:49Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"JUKDWZeQo8VBLkNf8tZARoK7KQS/BwK0NlSn3yj2eflPO+lcQAMoik/+KgLXK1wDB15JtBPiIMiaKMs8wwePAg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-08T09:09:38.279944Z"},"content_sha256":"b873db1b7504274c0d879f5196b05dd20c182dde4b3add316262ee905939823a","schema_version":"1.0","event_id":"sha256:b873db1b7504274c0d879f5196b05dd20c182dde4b3add316262ee905939823a"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:S2HT6JMPE6EJFHYIOC5MLY757P","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"ESLM: Risk-Averse Selective Language Modeling for Efficient Pretraining","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.LG","authors_text":"Melis Ilayda Bal, Michael Muehlebach, Volkan Cevher","submitted_at":"2025-05-26T12:23:26Z","abstract_excerpt":"Large language model pretraining is compute-intensive, yet many tokens contribute marginally to learning, resulting in inefficiency. We introduce Efficient Selective Language Modeling (ESLM), a risk-aware algorithm that improves training efficiency and distributional robustness by performing online token-level batch selection. ESLM leverages per-token statistics (e.g., entropy or loss) and applies value-at-risk thresholding to retain only the most informative tokens per batch. This data-centric mechanism reshapes the training loss, prioritizing high-risk tokens and eliminating redundant gradie"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.19893","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/2505.19893/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":"2026-07-05T11:09:49Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"QHsyRRFCXd2vZ9Op1Nf1Lzb+ojGI79dLp/MSB9QhOKefPFr9t66237bgRqt5oUHzwETGVVUsTTHRzqGeTLozCA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-08T09:09:38.280937Z"},"content_sha256":"a7cf62d2571c46eb9f08ef304339f273f28de1c16fea5b121d47a0d68d6a96f1","schema_version":"1.0","event_id":"sha256:a7cf62d2571c46eb9f08ef304339f273f28de1c16fea5b121d47a0d68d6a96f1"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/S2HT6JMPE6EJFHYIOC5MLY757P/bundle.json","state_url":"https://pith.science/pith/S2HT6JMPE6EJFHYIOC5MLY757P/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/S2HT6JMPE6EJFHYIOC5MLY757P/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":"2026-08-08T09:09:38Z","links":{"resolver":"https://pith.science/pith/S2HT6JMPE6EJFHYIOC5MLY757P","bundle":"https://pith.science/pith/S2HT6JMPE6EJFHYIOC5MLY757P/bundle.json","state":"https://pith.science/pith/S2HT6JMPE6EJFHYIOC5MLY757P/state.json","well_known_bundle":"https://pith.science/.well-known/pith/S2HT6JMPE6EJFHYIOC5MLY757P/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:S2HT6JMPE6EJFHYIOC5MLY757P","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":"84443c770689619c652869503566e6a8516d78ff83389a010597ac711381aab4","cross_cats_sorted":["cs.CL"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-05-26T12:23:26Z","title_canon_sha256":"283e93c2b269f37f0d34aeb04f1754f7f049de5899bc5675c43fa3f07a148691"},"schema_version":"1.0","source":{"id":"2505.19893","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2505.19893","created_at":"2026-07-05T11:09:49Z"},{"alias_kind":"arxiv_version","alias_value":"2505.19893v1","created_at":"2026-07-05T11:09:49Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.19893","created_at":"2026-07-05T11:09:49Z"},{"alias_kind":"pith_short_12","alias_value":"S2HT6JMPE6EJ","created_at":"2026-07-05T11:09:49Z"},{"alias_kind":"pith_short_16","alias_value":"S2HT6JMPE6EJFHYI","created_at":"2026-07-05T11:09:49Z"},{"alias_kind":"pith_short_8","alias_value":"S2HT6JMP","created_at":"2026-07-05T11:09:49Z"}],"graph_snapshots":[{"event_id":"sha256:a7cf62d2571c46eb9f08ef304339f273f28de1c16fea5b121d47a0d68d6a96f1","target":"graph","created_at":"2026-07-05T11:09:49Z","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/2505.19893/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Large language model pretraining is compute-intensive, yet many tokens contribute marginally to learning, resulting in inefficiency. We introduce Efficient Selective Language Modeling (ESLM), a risk-aware algorithm that improves training efficiency and distributional robustness by performing online token-level batch selection. ESLM leverages per-token statistics (e.g., entropy or loss) and applies value-at-risk thresholding to retain only the most informative tokens per batch. This data-centric mechanism reshapes the training loss, prioritizing high-risk tokens and eliminating redundant gradie","authors_text":"Melis Ilayda Bal, Michael Muehlebach, Volkan Cevher","cross_cats":["cs.CL"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-05-26T12:23:26Z","title":"ESLM: Risk-Averse Selective Language Modeling for Efficient Pretraining"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.19893","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:b873db1b7504274c0d879f5196b05dd20c182dde4b3add316262ee905939823a","target":"record","created_at":"2026-07-05T11:09:49Z","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":"84443c770689619c652869503566e6a8516d78ff83389a010597ac711381aab4","cross_cats_sorted":["cs.CL"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-05-26T12:23:26Z","title_canon_sha256":"283e93c2b269f37f0d34aeb04f1754f7f049de5899bc5675c43fa3f07a148691"},"schema_version":"1.0","source":{"id":"2505.19893","kind":"arxiv","version":1}},"canonical_sha256":"968f3f258f2788929f0870bac5e3fdfbe1cc74171d0c94ec105959d507e2acc7","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"968f3f258f2788929f0870bac5e3fdfbe1cc74171d0c94ec105959d507e2acc7","first_computed_at":"2026-07-05T11:09:49.569024Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:09:49.569024Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"YPBEGj9HVuwyLDcDJEAgR3ZUIq6WDIJ6IxhxbD7aifjQATlLWQkTY1P+BagZBpQkJF0GTT4oJ7PGxFs2Iru3Aw==","signature_status":"signed_v1","signed_at":"2026-07-05T11:09:49.569534Z","signed_message":"canonical_sha256_bytes"},"source_id":"2505.19893","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:b873db1b7504274c0d879f5196b05dd20c182dde4b3add316262ee905939823a","sha256:a7cf62d2571c46eb9f08ef304339f273f28de1c16fea5b121d47a0d68d6a96f1"],"state_sha256":"3a8eda4ae4e85c6c69c895b8b718a352331fac2bd6ad39c6539120284736e4d0"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"4D0liT9srA+1uB/+wQrTmOTYVVwcsDP796zGLo3JnfxIEMJqxQVNb0UmIigoq3d0YJWivC8y5y4NGR4UiYXsCA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-08T09:09:38.288703Z","bundle_sha256":"c9896f710238fefc2a4d16303de21651c8cdd669b77aae34c60e4eee8ba92f50"}}