{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2023:CAI4IMOSTWORB26XEK3HMLNKKA","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":"fc96d70f4339ef6a7ea4f91c0a84441b8dc996a69d85a3dd76178a6f181d107c","cross_cats_sorted":["cs.LG"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-11-28T03:23:20Z","title_canon_sha256":"af5df5ac73aa5e580bcc05a69a919e9f79e524916c2b8436161eafa0c94e374d"},"schema_version":"1.0","source":{"id":"2312.03732","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2312.03732","created_at":"2026-05-17T23:38:50Z"},{"alias_kind":"arxiv_version","alias_value":"2312.03732v1","created_at":"2026-05-17T23:38:50Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2312.03732","created_at":"2026-05-17T23:38:50Z"},{"alias_kind":"pith_short_12","alias_value":"CAI4IMOSTWOR","created_at":"2026-05-18T12:33:33Z"},{"alias_kind":"pith_short_16","alias_value":"CAI4IMOSTWORB26X","created_at":"2026-05-18T12:33:33Z"},{"alias_kind":"pith_short_8","alias_value":"CAI4IMOS","created_at":"2026-05-18T12:33:33Z"}],"graph_snapshots":[{"event_id":"sha256:4514914d460528bf3ac4531559f6ed1a9b3bfb625164ccb9d3cce51ccdc918fa","target":"graph","created_at":"2026-05-17T23:38:50Z","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":"we study the impact of the scaling factor on the learning process and prove that LoRA adapters should be divided by a factor of the square root of the rank"},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"The proof that 1/sqrt(rank) is optimal rests on unstated assumptions about initialization variance, gradient flow, and the precise form of the LoRA update rule during fine-tuning; these assumptions are not detailed in the provided abstract."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"LoRA adapters should be scaled by 1/sqrt(rank) rather than 1/rank to stabilize learning and enable effective use of higher ranks during fine-tuning of large language models."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"LoRA adapters should be scaled by dividing by the square root of the rank rather than the full rank to stabilize learning."}],"snapshot_sha256":"228ad1f7685270356943d91311780a58336dccb02bf917c3b10a87eac8f2f96d"},"formal_canon":{"evidence_count":2,"snapshot_sha256":"1fded5ab56adef58140b9f279197ddf9fd002341cffcd2bd1c2d972b2a474ac4"},"paper":{"abstract_excerpt":"As large language models (LLMs) have become increasingly compute and memory intensive, parameter-efficient fine-tuning (PEFT) methods are now a common strategy to fine-tune LLMs. A popular PEFT method is Low-Rank Adapters (LoRA), which adds trainable low-rank \"adapters\" to selected layers. Each adapter consists of a low-rank matrix product, multiplicatively scaled by a rank-dependent factor. This scaling factor, which divides adapters by a factor of the rank, results in slowed learning and stunted performance for LoRA with higher-rank adapters. Consequently, the use of LoRA in practice has gen","authors_text":"Damjan Kalajdzievski","cross_cats":["cs.LG"],"headline":"LoRA adapters should be scaled by dividing by the square root of the rank rather than the full rank to stabilize learning.","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-11-28T03:23:20Z","title":"A Rank Stabilization Scaling Factor for Fine-Tuning with LoRA"},"references":{"count":91,"internal_anchors":6,"resolved_work":91,"sample":[{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":1,"title":"2023 , publisher =","work_id":"4e8b0b95-ae6c-4edf-8218-6b1fbbfd9734","year":2023},{"cited_arxiv_id":"","doi":"10.5281/zenodo.5371628","is_internal_anchor":false,"ref_index":2,"title":"Aaron Gokaslan, Vanya Cohen, Ellie Pavlick, and Stefanie Tellex","work_id":"27d00efc-1c8b-4af4-bd30-5dfa575d0989","year":null},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":3,"title":"Think you have Solved Question Answering? Try ARC, the AI2 Reasoning Challenge , author=. 2018 , eprint=","work_id":"9880c11f-6c71-4e36-9736-affdf3d20344","year":2018},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":4,"title":"HellaSwag: Can a Machine Really Finish Your Sentence? , author=. 2019 , eprint=","work_id":"b5234106-7b98-47af-87c3-45d428066901","year":2019},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":5,"title":"Measuring Massive Multitask Language Understanding , author=. 2021 , eprint=","work_id":"01eb26fa-b178-499b-bc7b-b2ff5dd67b24","year":2021}],"snapshot_sha256":"fea4b5c0f7a8f09e836708a7210ddf6567efda57631d4dc9edc69c6adae48df3"},"source":{"id":"2312.03732","kind":"arxiv","version":1},"verdict":{"created_at":"2026-05-15T22:02:04.412118Z","id":"c9b1eb0e-c4c8-43ef-a1df-9e713c8a8353","model_set":{"reader":"grok-4.3"},"one_line_summary":"LoRA adapters should be scaled by 1/sqrt(rank) rather than 1/rank to stabilize learning and enable effective use of higher ranks during fine-tuning of large language models.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"LoRA adapters should be scaled by dividing by the square root of the rank rather than the full rank to stabilize learning.","strongest_claim":"we study the impact of the scaling factor on the learning process and prove that LoRA adapters should be divided by a factor of the square root of the rank","weakest_assumption":"The proof that 1/sqrt(rank) is optimal rests on unstated assumptions about initialization variance, gradient flow, and the precise form of the LoRA update rule during fine-tuning; these assumptions are not detailed in the provided abstract."}},"verdict_id":"c9b1eb0e-c4c8-43ef-a1df-9e713c8a8353"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:81cc643860f9235ece695740400b29a15275c0cf39113f551db4c797eec9c0a0","target":"record","created_at":"2026-05-17T23:38:50Z","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":"fc96d70f4339ef6a7ea4f91c0a84441b8dc996a69d85a3dd76178a6f181d107c","cross_cats_sorted":["cs.LG"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-11-28T03:23:20Z","title_canon_sha256":"af5df5ac73aa5e580bcc05a69a919e9f79e524916c2b8436161eafa0c94e374d"},"schema_version":"1.0","source":{"id":"2312.03732","kind":"arxiv","version":1}},"canonical_sha256":"1011c431d29d9d10ebd722b6762daa50049629f32c1a69c94a3fd0ed1f29b933","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"1011c431d29d9d10ebd722b6762daa50049629f32c1a69c94a3fd0ed1f29b933","first_computed_at":"2026-05-17T23:38:50.002668Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-17T23:38:50.002668Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"UqMXcBRQltHaEhdzJWNF3thAJXl7P1sLwmO2B5ZY3z8fxiwyCnct463CWYxDFYTOtNSc43XnX/w1qlUnPApWDQ==","signature_status":"signed_v1","signed_at":"2026-05-17T23:38:50.003179Z","signed_message":"canonical_sha256_bytes"},"source_id":"2312.03732","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:81cc643860f9235ece695740400b29a15275c0cf39113f551db4c797eec9c0a0","sha256:4514914d460528bf3ac4531559f6ed1a9b3bfb625164ccb9d3cce51ccdc918fa"],"state_sha256":"1da607a4b0794fb16d132ab3c1398f580abedfbb2def527d65ec34b6d39cf47c"}