{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:2S6NTSNWQHHKBUUHZG4Z4DTZIH","short_pith_number":"pith:2S6NTSNW","canonical_record":{"source":{"id":"2605.01046","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-01T19:20:25Z","cross_cats_sorted":[],"title_canon_sha256":"8455f47f3b3d9eb693aaa048a1d58d4bca23a47090c0e4f76cd76b736325e538","abstract_canon_sha256":"8f6c5a4e907da9540ed1f1c0d05b1ee17aaa2c86cde169671a699ebe5d5535c9"},"schema_version":"1.0"},"canonical_sha256":"d4bcd9c9b681cea0d287c9b99e0e7941c44117f5d574d3b87a011793317aa04e","source":{"kind":"arxiv","id":"2605.01046","version":3},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.01046","created_at":"2026-05-28T00:05:15Z"},{"alias_kind":"arxiv_version","alias_value":"2605.01046v3","created_at":"2026-05-28T00:05:15Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.01046","created_at":"2026-05-28T00:05:15Z"},{"alias_kind":"pith_short_12","alias_value":"2S6NTSNWQHHK","created_at":"2026-05-28T00:05:15Z"},{"alias_kind":"pith_short_16","alias_value":"2S6NTSNWQHHKBUUH","created_at":"2026-05-28T00:05:15Z"},{"alias_kind":"pith_short_8","alias_value":"2S6NTSNW","created_at":"2026-05-28T00:05:15Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:2S6NTSNWQHHKBUUHZG4Z4DTZIH","target":"record","payload":{"canonical_record":{"source":{"id":"2605.01046","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-01T19:20:25Z","cross_cats_sorted":[],"title_canon_sha256":"8455f47f3b3d9eb693aaa048a1d58d4bca23a47090c0e4f76cd76b736325e538","abstract_canon_sha256":"8f6c5a4e907da9540ed1f1c0d05b1ee17aaa2c86cde169671a699ebe5d5535c9"},"schema_version":"1.0"},"canonical_sha256":"d4bcd9c9b681cea0d287c9b99e0e7941c44117f5d574d3b87a011793317aa04e","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-28T00:05:15.212793Z","signature_b64":"XuxKh0f/cRo5ZlVvBrTcgKM3Y3//x1pm1GBjngbIR5if5XeOrDlJ6Ase2twAP33BVTLmkK39OP+9/vPeyTJgBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"d4bcd9c9b681cea0d287c9b99e0e7941c44117f5d574d3b87a011793317aa04e","last_reissued_at":"2026-05-28T00:05:15.211835Z","signature_status":"signed_v1","first_computed_at":"2026-05-28T00:05:15.211835Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2605.01046","source_version":3,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-28T00:05:15Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"xM9UA81+PaW6S8cO8Z2Xp0335GoncbgUflh4qQmBoO3EMXMQD0trZPY0LONHMWXPYsK5+H9fxl8jRGUMMtoIAg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-06T18:57:11.071291Z"},"content_sha256":"51fb9373ae4d3817ad7e29e62442339e7b6661ab405b54fe610f3612263f9784","schema_version":"1.0","event_id":"sha256:51fb9373ae4d3817ad7e29e62442339e7b6661ab405b54fe610f3612263f9784"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:2S6NTSNWQHHKBUUHZG4Z4DTZIH","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Learning in the Fisher Subspace: A Guided Initialization for LoRA Fine-Tuning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"Using Fisher curvature from downstream data to initialize LoRA subspaces improves fine-tuning performance over weight-only methods.","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Hung-Yu Kao, Ying-Jia Lin, Zhi-Quan Feng","submitted_at":"2026-05-01T19:20:25Z","abstract_excerpt":"LoRA adapts large language models (LLMs) by restricting updates to low-rank subspaces of pre-trained weights. While this substantially reduces training cost, the effectiveness of adaptation critically depends on which subspace is chosen at initialization: a poor initialization that allocates capacity to task-irrelevant directions can severely hinder downstream performance. Existing initialization strategies primarily rely on the intrinsic properties of pre-trained weights, implicitly assuming that weight geometry alone reflects task relevance. However, such criteria overlook how the model inte"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Empirical results across diverse tasks and modalities demonstrate that data-aware initialization consistently and significantly improves downstream performance over existing approaches.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That the curvature information induced by the downstream data distribution accurately identifies parameter directions whose perturbations most influence model predictions on the target objective.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"Fisher information from target data provides a better criterion than weight geometry for choosing LoRA subspaces, yielding consistent performance gains on downstream tasks.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Using Fisher curvature from downstream data to initialize LoRA subspaces improves fine-tuning performance over weight-only methods.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"4fe263de3d7f2221f758ef3207f7a5448cd5070aafd6700689665855ab8d5c6e"},"source":{"id":"2605.01046","kind":"arxiv","version":3},"verdict":{"id":"95a153b7-20d6-4c4b-97f8-c6e2655c7e5d","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-09T19:30:10.890636Z","strongest_claim":"Empirical results across diverse tasks and modalities demonstrate that data-aware initialization consistently and significantly improves downstream performance over existing approaches.","one_line_summary":"Fisher information from target data provides a better criterion than weight geometry for choosing LoRA subspaces, yielding consistent performance gains on downstream tasks.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That the curvature information induced by the downstream data distribution accurately identifies parameter directions whose perturbations most influence model predictions on the target objective.","pith_extraction_headline":"Using Fisher curvature from downstream data to initialize LoRA subspaces improves fine-tuning performance over weight-only methods."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2605.01046/integrity.json","findings":[],"available":true,"detectors_run":[{"name":"ai_meta_artifact","ran_at":"2026-05-20T18:39:26.354215Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_compliance","ran_at":"2026-05-19T17:40:54.459980Z","status":"completed","version":"1.0.0","findings_count":0}],"snapshot_sha256":"a18f734de0fb534ad988031bd386a8993ac4714d14d0fd0bf72e76c0f01ca21b"},"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":"95a153b7-20d6-4c4b-97f8-c6e2655c7e5d"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-28T00:05:15Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"cpt98gpji15qN8fZ0gOoJBqwxbfmOsOhN2xFsPS3VPlPh+cQ+VSLZVs7+EIMmgGL3mt+4K30KF610T2xp+0kDw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-06T18:57:11.071838Z"},"content_sha256":"c6ecda1edd12c965f0d7e6196ac9291136dd52fc71144844e1ec7be99348940b","schema_version":"1.0","event_id":"sha256:c6ecda1edd12c965f0d7e6196ac9291136dd52fc71144844e1ec7be99348940b"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/2S6NTSNWQHHKBUUHZG4Z4DTZIH/bundle.json","state_url":"https://pith.science/pith/2S6NTSNWQHHKBUUHZG4Z4DTZIH/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/2S6NTSNWQHHKBUUHZG4Z4DTZIH/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":"2026-06-06T18:57:11Z","links":{"resolver":"https://pith.science/pith/2S6NTSNWQHHKBUUHZG4Z4DTZIH","bundle":"https://pith.science/pith/2S6NTSNWQHHKBUUHZG4Z4DTZIH/bundle.json","state":"https://pith.science/pith/2S6NTSNWQHHKBUUHZG4Z4DTZIH/state.json","well_known_bundle":"https://pith.science/.well-known/pith/2S6NTSNWQHHKBUUHZG4Z4DTZIH/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:2S6NTSNWQHHKBUUHZG4Z4DTZIH","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":"8f6c5a4e907da9540ed1f1c0d05b1ee17aaa2c86cde169671a699ebe5d5535c9","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-01T19:20:25Z","title_canon_sha256":"8455f47f3b3d9eb693aaa048a1d58d4bca23a47090c0e4f76cd76b736325e538"},"schema_version":"1.0","source":{"id":"2605.01046","kind":"arxiv","version":3}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.01046","created_at":"2026-05-28T00:05:15Z"},{"alias_kind":"arxiv_version","alias_value":"2605.01046v3","created_at":"2026-05-28T00:05:15Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.01046","created_at":"2026-05-28T00:05:15Z"},{"alias_kind":"pith_short_12","alias_value":"2S6NTSNWQHHK","created_at":"2026-05-28T00:05:15Z"},{"alias_kind":"pith_short_16","alias_value":"2S6NTSNWQHHKBUUH","created_at":"2026-05-28T00:05:15Z"},{"alias_kind":"pith_short_8","alias_value":"2S6NTSNW","created_at":"2026-05-28T00:05:15Z"}],"graph_snapshots":[{"event_id":"sha256:c6ecda1edd12c965f0d7e6196ac9291136dd52fc71144844e1ec7be99348940b","target":"graph","created_at":"2026-05-28T00:05:15Z","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":4,"items":[{"attestation":"unclaimed","claim_id":"C1","kind":"strongest_claim","source":"verdict.strongest_claim","status":"machine_extracted","text":"Empirical results across diverse tasks and modalities demonstrate that data-aware initialization consistently and significantly improves downstream performance over existing approaches."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"That the curvature information induced by the downstream data distribution accurately identifies parameter directions whose perturbations most influence model predictions on the target objective."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"Fisher information from target data provides a better criterion than weight geometry for choosing LoRA subspaces, yielding consistent performance gains on downstream tasks."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"Using Fisher curvature from downstream data to initialize LoRA subspaces improves fine-tuning performance over weight-only methods."}],"snapshot_sha256":"4fe263de3d7f2221f758ef3207f7a5448cd5070aafd6700689665855ab8d5c6e"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[{"findings_count":0,"name":"ai_meta_artifact","ran_at":"2026-05-20T18:39:26.354215Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"doi_compliance","ran_at":"2026-05-19T17:40:54.459980Z","status":"completed","version":"1.0.0"}],"endpoint":"/pith/2605.01046/integrity.json","findings":[],"snapshot_sha256":"a18f734de0fb534ad988031bd386a8993ac4714d14d0fd0bf72e76c0f01ca21b","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"LoRA adapts large language models (LLMs) by restricting updates to low-rank subspaces of pre-trained weights. While this substantially reduces training cost, the effectiveness of adaptation critically depends on which subspace is chosen at initialization: a poor initialization that allocates capacity to task-irrelevant directions can severely hinder downstream performance. Existing initialization strategies primarily rely on the intrinsic properties of pre-trained weights, implicitly assuming that weight geometry alone reflects task relevance. However, such criteria overlook how the model inte","authors_text":"Hung-Yu Kao, Ying-Jia Lin, Zhi-Quan Feng","cross_cats":[],"headline":"Using Fisher curvature from downstream data to initialize LoRA subspaces improves fine-tuning performance over weight-only methods.","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-01T19:20:25Z","title":"Learning in the Fisher Subspace: A Guided Initialization for LoRA Fine-Tuning"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.01046","kind":"arxiv","version":3},"verdict":{"created_at":"2026-05-09T19:30:10.890636Z","id":"95a153b7-20d6-4c4b-97f8-c6e2655c7e5d","model_set":{"reader":"grok-4.3"},"one_line_summary":"Fisher information from target data provides a better criterion than weight geometry for choosing LoRA subspaces, yielding consistent performance gains on downstream tasks.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"Using Fisher curvature from downstream data to initialize LoRA subspaces improves fine-tuning performance over weight-only methods.","strongest_claim":"Empirical results across diverse tasks and modalities demonstrate that data-aware initialization consistently and significantly improves downstream performance over existing approaches.","weakest_assumption":"That the curvature information induced by the downstream data distribution accurately identifies parameter directions whose perturbations most influence model predictions on the target objective."}},"verdict_id":"95a153b7-20d6-4c4b-97f8-c6e2655c7e5d"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:51fb9373ae4d3817ad7e29e62442339e7b6661ab405b54fe610f3612263f9784","target":"record","created_at":"2026-05-28T00:05:15Z","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":"8f6c5a4e907da9540ed1f1c0d05b1ee17aaa2c86cde169671a699ebe5d5535c9","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-01T19:20:25Z","title_canon_sha256":"8455f47f3b3d9eb693aaa048a1d58d4bca23a47090c0e4f76cd76b736325e538"},"schema_version":"1.0","source":{"id":"2605.01046","kind":"arxiv","version":3}},"canonical_sha256":"d4bcd9c9b681cea0d287c9b99e0e7941c44117f5d574d3b87a011793317aa04e","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"d4bcd9c9b681cea0d287c9b99e0e7941c44117f5d574d3b87a011793317aa04e","first_computed_at":"2026-05-28T00:05:15.211835Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-28T00:05:15.211835Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"XuxKh0f/cRo5ZlVvBrTcgKM3Y3//x1pm1GBjngbIR5if5XeOrDlJ6Ase2twAP33BVTLmkK39OP+9/vPeyTJgBg==","signature_status":"signed_v1","signed_at":"2026-05-28T00:05:15.212793Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.01046","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:51fb9373ae4d3817ad7e29e62442339e7b6661ab405b54fe610f3612263f9784","sha256:c6ecda1edd12c965f0d7e6196ac9291136dd52fc71144844e1ec7be99348940b"],"state_sha256":"27d952ae6a9a523c3723c2dbe18c1fec6b2ff0cbf670288e1c694f492e7efa12"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"NrqrJByoOkW50k5agPUMil5NHIwDC4FNb1h/qv3cu02ik4SSP6Ch9/k3tJf1Zn0Gn4nm0c+YBoKExhte9CCSDg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-06-06T18:57:11.074702Z","bundle_sha256":"c1e6352a58253a35d72a2d3f7acd465da889b93754fd2a35bc9dd903db852728"}}