{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:3YQV2UD7ROXTKXACCPNFK3V5B7","short_pith_number":"pith:3YQV2UD7","canonical_record":{"source":{"id":"2406.06623","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-06-07T21:20:57Z","cross_cats_sorted":["stat.ML"],"title_canon_sha256":"6c263cf6ba39abc1fcbda66006575c4375668b9bc5fa7f3da5c0e67266c2b041","abstract_canon_sha256":"3076627e0496450d4406ae595101c138d6f6c7e2ad5e286d48a55e5e990f4282"},"schema_version":"1.0"},"canonical_sha256":"de215d507f8baf355c0213da556ebd0fd0321394874798ce3f641cf7610c31e5","source":{"kind":"arxiv","id":"2406.06623","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2406.06623","created_at":"2026-07-05T08:29:48Z"},{"alias_kind":"arxiv_version","alias_value":"2406.06623v1","created_at":"2026-07-05T08:29:48Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2406.06623","created_at":"2026-07-05T08:29:48Z"},{"alias_kind":"pith_short_12","alias_value":"3YQV2UD7ROXT","created_at":"2026-07-05T08:29:48Z"},{"alias_kind":"pith_short_16","alias_value":"3YQV2UD7ROXTKXAC","created_at":"2026-07-05T08:29:48Z"},{"alias_kind":"pith_short_8","alias_value":"3YQV2UD7","created_at":"2026-07-05T08:29:48Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:3YQV2UD7ROXTKXACCPNFK3V5B7","target":"record","payload":{"canonical_record":{"source":{"id":"2406.06623","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-06-07T21:20:57Z","cross_cats_sorted":["stat.ML"],"title_canon_sha256":"6c263cf6ba39abc1fcbda66006575c4375668b9bc5fa7f3da5c0e67266c2b041","abstract_canon_sha256":"3076627e0496450d4406ae595101c138d6f6c7e2ad5e286d48a55e5e990f4282"},"schema_version":"1.0"},"canonical_sha256":"de215d507f8baf355c0213da556ebd0fd0321394874798ce3f641cf7610c31e5","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:29:48.248412Z","signature_b64":"sTEy2mX4EsOOFwP7+3LxpeAU2latHAYD1uoZjUniQLvcF2MZj3UCpFwjRH8/MSFmK0b7H6+6lUF4IGnumnDxAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"de215d507f8baf355c0213da556ebd0fd0321394874798ce3f641cf7610c31e5","last_reissued_at":"2026-07-05T08:29:48.248044Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:29:48.248044Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2406.06623","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":"2026-07-05T08:29:48Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"4v64uH/8k+m6yPMxQHj0cxkcCbfdOfFzJnseBgCdIlih1Vp13mfcMPTRPFbngCWzcOVIxD0hS+FPkuFKVpZhDg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-20T17:44:51.110672Z"},"content_sha256":"c29439fd61633f4e0fd00b3f86b521f02203405f38030eee12b20d45ad43c5e8","schema_version":"1.0","event_id":"sha256:c29439fd61633f4e0fd00b3f86b521f02203405f38030eee12b20d45ad43c5e8"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:3YQV2UD7ROXTKXACCPNFK3V5B7","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Spectrum: Targeted Training on Signal to Noise Ratio","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"David Golchinfar, Eric Hartford, Fernando Fernandes Neto, Lucas Atkins","submitted_at":"2024-06-07T21:20:57Z","abstract_excerpt":"Efficiently post-training large language models remains a challenging task due to the vast computational resources required. We present Spectrum, a method that accelerates LLM training by selectively targeting layer modules based on their signal-to-noise ratio (SNR), and freezing the remaining modules. Our approach, which utilizes an algorithm to compute module SNRs prior to training, has shown to effectively match the performance of full fine-tuning while reducing GPU memory usage. Experiments comparing Spectrum to existing methods such as QLoRA demonstrate its effectiveness in terms of model"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2406.06623","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/2406.06623/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":"2026-07-05T08:29:48Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"EOoabsUR2SvACuDybyAIL9wsnLqvV/SEFTRQ1WxxlOXFI50Bot4+eFaDOj5FuxDmnTl5qqO0/mEOzbhdkqvxCQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-20T17:44:51.111031Z"},"content_sha256":"f732c2991b640f9e59b5b06d502abb661b87c15ce7ca634daaeb75ee71bd571b","schema_version":"1.0","event_id":"sha256:f732c2991b640f9e59b5b06d502abb661b87c15ce7ca634daaeb75ee71bd571b"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/3YQV2UD7ROXTKXACCPNFK3V5B7/bundle.json","state_url":"https://pith.science/pith/3YQV2UD7ROXTKXACCPNFK3V5B7/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/3YQV2UD7ROXTKXACCPNFK3V5B7/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":"2026-08-20T17:44:51Z","links":{"resolver":"https://pith.science/pith/3YQV2UD7ROXTKXACCPNFK3V5B7","bundle":"https://pith.science/pith/3YQV2UD7ROXTKXACCPNFK3V5B7/bundle.json","state":"https://pith.science/pith/3YQV2UD7ROXTKXACCPNFK3V5B7/state.json","well_known_bundle":"https://pith.science/.well-known/pith/3YQV2UD7ROXTKXACCPNFK3V5B7/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:3YQV2UD7ROXTKXACCPNFK3V5B7","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":"3076627e0496450d4406ae595101c138d6f6c7e2ad5e286d48a55e5e990f4282","cross_cats_sorted":["stat.ML"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-06-07T21:20:57Z","title_canon_sha256":"6c263cf6ba39abc1fcbda66006575c4375668b9bc5fa7f3da5c0e67266c2b041"},"schema_version":"1.0","source":{"id":"2406.06623","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2406.06623","created_at":"2026-07-05T08:29:48Z"},{"alias_kind":"arxiv_version","alias_value":"2406.06623v1","created_at":"2026-07-05T08:29:48Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2406.06623","created_at":"2026-07-05T08:29:48Z"},{"alias_kind":"pith_short_12","alias_value":"3YQV2UD7ROXT","created_at":"2026-07-05T08:29:48Z"},{"alias_kind":"pith_short_16","alias_value":"3YQV2UD7ROXTKXAC","created_at":"2026-07-05T08:29:48Z"},{"alias_kind":"pith_short_8","alias_value":"3YQV2UD7","created_at":"2026-07-05T08:29:48Z"}],"graph_snapshots":[{"event_id":"sha256:f732c2991b640f9e59b5b06d502abb661b87c15ce7ca634daaeb75ee71bd571b","target":"graph","created_at":"2026-07-05T08:29:48Z","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/2406.06623/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Efficiently post-training large language models remains a challenging task due to the vast computational resources required. We present Spectrum, a method that accelerates LLM training by selectively targeting layer modules based on their signal-to-noise ratio (SNR), and freezing the remaining modules. Our approach, which utilizes an algorithm to compute module SNRs prior to training, has shown to effectively match the performance of full fine-tuning while reducing GPU memory usage. Experiments comparing Spectrum to existing methods such as QLoRA demonstrate its effectiveness in terms of model","authors_text":"David Golchinfar, Eric Hartford, Fernando Fernandes Neto, Lucas Atkins","cross_cats":["stat.ML"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-06-07T21:20:57Z","title":"Spectrum: Targeted Training on Signal to Noise Ratio"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2406.06623","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:c29439fd61633f4e0fd00b3f86b521f02203405f38030eee12b20d45ad43c5e8","target":"record","created_at":"2026-07-05T08:29:48Z","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":"3076627e0496450d4406ae595101c138d6f6c7e2ad5e286d48a55e5e990f4282","cross_cats_sorted":["stat.ML"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-06-07T21:20:57Z","title_canon_sha256":"6c263cf6ba39abc1fcbda66006575c4375668b9bc5fa7f3da5c0e67266c2b041"},"schema_version":"1.0","source":{"id":"2406.06623","kind":"arxiv","version":1}},"canonical_sha256":"de215d507f8baf355c0213da556ebd0fd0321394874798ce3f641cf7610c31e5","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"de215d507f8baf355c0213da556ebd0fd0321394874798ce3f641cf7610c31e5","first_computed_at":"2026-07-05T08:29:48.248044Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T08:29:48.248044Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"sTEy2mX4EsOOFwP7+3LxpeAU2latHAYD1uoZjUniQLvcF2MZj3UCpFwjRH8/MSFmK0b7H6+6lUF4IGnumnDxAw==","signature_status":"signed_v1","signed_at":"2026-07-05T08:29:48.248412Z","signed_message":"canonical_sha256_bytes"},"source_id":"2406.06623","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:c29439fd61633f4e0fd00b3f86b521f02203405f38030eee12b20d45ad43c5e8","sha256:f732c2991b640f9e59b5b06d502abb661b87c15ce7ca634daaeb75ee71bd571b"],"state_sha256":"e34c6d4f65a2ad668bd3700494f5f5d91f614d7c1a40c4e724825eb14e12c6a7"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"Jx375ynyJqrmkGHf6U7iYQIbeLW4IAqWu0c+k8bTFOmmUu3OIXIF+PhQ4zBjbm2nmivjKPt+ZN/R5oEsaABeBA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-20T17:44:51.114173Z","bundle_sha256":"37faa5f5d4d2785fe9a6f1ce3a8261e13cfb528279f98518674d2cb555d36e33"}}