{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:KJK65QQKIH55ZEUFFEDM57J7RI","short_pith_number":"pith:KJK65QQK","canonical_record":{"source":{"id":"2406.11753","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-06-17T17:13:08Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"b53e69677dd2f001872252684ee15e15817327f405fe48d5d16c0bd6e3855296","abstract_canon_sha256":"368bff7f68d2dff435b93b21020696655d6f1e51b7b5aef2265ff142ee679b25"},"schema_version":"1.0"},"canonical_sha256":"5255eec20a41fbdc92852906cefd3f8a2474a63b6c1873373b49fa59144cca55","source":{"kind":"arxiv","id":"2406.11753","version":3},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2406.11753","created_at":"2026-07-05T11:13:10Z"},{"alias_kind":"arxiv_version","alias_value":"2406.11753v3","created_at":"2026-07-05T11:13:10Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2406.11753","created_at":"2026-07-05T11:13:10Z"},{"alias_kind":"pith_short_12","alias_value":"KJK65QQKIH55","created_at":"2026-07-05T11:13:10Z"},{"alias_kind":"pith_short_16","alias_value":"KJK65QQKIH55ZEUF","created_at":"2026-07-05T11:13:10Z"},{"alias_kind":"pith_short_8","alias_value":"KJK65QQK","created_at":"2026-07-05T11:13:10Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:KJK65QQKIH55ZEUFFEDM57J7RI","target":"record","payload":{"canonical_record":{"source":{"id":"2406.11753","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-06-17T17:13:08Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"b53e69677dd2f001872252684ee15e15817327f405fe48d5d16c0bd6e3855296","abstract_canon_sha256":"368bff7f68d2dff435b93b21020696655d6f1e51b7b5aef2265ff142ee679b25"},"schema_version":"1.0"},"canonical_sha256":"5255eec20a41fbdc92852906cefd3f8a2474a63b6c1873373b49fa59144cca55","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:13:10.709332Z","signature_b64":"BUmMS6oQrK8IlpXENzdS3JDcZM96HsL3EdBX3MCXoFoYF7Xz6WBe5gjBXv/E5ACMQwc7VTDvt0OWKimPPhyLBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"5255eec20a41fbdc92852906cefd3f8a2474a63b6c1873373b49fa59144cca55","last_reissued_at":"2026-07-05T11:13:10.708812Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:13:10.708812Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2406.11753","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":"2026-07-05T11:13:10Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"6wxwDI9/q8XGZRuCPP36uisG4wYcl8Omk24DTJIqt6lmL37PlgBYXDWn2+Nzirt5djTRS7TSbdkHa7Ddy++AAw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-09T15:30:00.067759Z"},"content_sha256":"21ccb1577584a0cdbf95fc895abcfd5c78985c03a1b407be613eb502d3d18ab1","schema_version":"1.0","event_id":"sha256:21ccb1577584a0cdbf95fc895abcfd5c78985c03a1b407be613eb502d3d18ab1"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:KJK65QQKIH55ZEUFFEDM57J7RI","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"A Semantic-Aware Layer-Freezing Approach to Computation-Efficient Fine-Tuning of Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Aldeida Aleti, Chunyang Chen, Hongyu Zhang, Jian Gu","submitted_at":"2024-06-17T17:13:08Z","abstract_excerpt":"Finetuning language models (LMs) is crucial for adapting the models to downstream data and tasks. However, full finetuning is usually costly. Existing work, such as parameter-efficient finetuning (PEFT), often focuses on \\textit{how to finetune} but neglects the issue of \\textit{where to finetune}. As a pioneering work on reducing the cost of backpropagation (at the layer level) by answering where to finetune, we conduct a semantic analysis of the LM inference process. We first propose using transition traces of the latent representation to compute deviations (or loss). Then, using a derived f"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2406.11753","kind":"arxiv","version":3},"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/2406.11753/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":"2026-07-05T11:13:10Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"2D40d1wuQ6pc459RtUKjIdTCg9nlrlC7Cvm7OYuZ1wPALOiWmgIXQHDL3qElrbBCsZy3elWoE014mh/hFoVyCA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-09T15:30:00.068325Z"},"content_sha256":"461fb1defe8ae2466ec77e749f93c61318927c1cf61c3ba9faa3900b0a775dc5","schema_version":"1.0","event_id":"sha256:461fb1defe8ae2466ec77e749f93c61318927c1cf61c3ba9faa3900b0a775dc5"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/KJK65QQKIH55ZEUFFEDM57J7RI/bundle.json","state_url":"https://pith.science/pith/KJK65QQKIH55ZEUFFEDM57J7RI/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/KJK65QQKIH55ZEUFFEDM57J7RI/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":"2026-08-09T15:30:00Z","links":{"resolver":"https://pith.science/pith/KJK65QQKIH55ZEUFFEDM57J7RI","bundle":"https://pith.science/pith/KJK65QQKIH55ZEUFFEDM57J7RI/bundle.json","state":"https://pith.science/pith/KJK65QQKIH55ZEUFFEDM57J7RI/state.json","well_known_bundle":"https://pith.science/.well-known/pith/KJK65QQKIH55ZEUFFEDM57J7RI/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:KJK65QQKIH55ZEUFFEDM57J7RI","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":"368bff7f68d2dff435b93b21020696655d6f1e51b7b5aef2265ff142ee679b25","cross_cats_sorted":["cs.LG"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-06-17T17:13:08Z","title_canon_sha256":"b53e69677dd2f001872252684ee15e15817327f405fe48d5d16c0bd6e3855296"},"schema_version":"1.0","source":{"id":"2406.11753","kind":"arxiv","version":3}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2406.11753","created_at":"2026-07-05T11:13:10Z"},{"alias_kind":"arxiv_version","alias_value":"2406.11753v3","created_at":"2026-07-05T11:13:10Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2406.11753","created_at":"2026-07-05T11:13:10Z"},{"alias_kind":"pith_short_12","alias_value":"KJK65QQKIH55","created_at":"2026-07-05T11:13:10Z"},{"alias_kind":"pith_short_16","alias_value":"KJK65QQKIH55ZEUF","created_at":"2026-07-05T11:13:10Z"},{"alias_kind":"pith_short_8","alias_value":"KJK65QQK","created_at":"2026-07-05T11:13:10Z"}],"graph_snapshots":[{"event_id":"sha256:461fb1defe8ae2466ec77e749f93c61318927c1cf61c3ba9faa3900b0a775dc5","target":"graph","created_at":"2026-07-05T11:13:10Z","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/2406.11753/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Finetuning language models (LMs) is crucial for adapting the models to downstream data and tasks. However, full finetuning is usually costly. Existing work, such as parameter-efficient finetuning (PEFT), often focuses on \\textit{how to finetune} but neglects the issue of \\textit{where to finetune}. As a pioneering work on reducing the cost of backpropagation (at the layer level) by answering where to finetune, we conduct a semantic analysis of the LM inference process. We first propose using transition traces of the latent representation to compute deviations (or loss). Then, using a derived f","authors_text":"Aldeida Aleti, Chunyang Chen, Hongyu Zhang, Jian Gu","cross_cats":["cs.LG"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-06-17T17:13:08Z","title":"A Semantic-Aware Layer-Freezing Approach to Computation-Efficient Fine-Tuning of Language Models"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2406.11753","kind":"arxiv","version":3},"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":"sha256:21ccb1577584a0cdbf95fc895abcfd5c78985c03a1b407be613eb502d3d18ab1","target":"record","created_at":"2026-07-05T11:13:10Z","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":"368bff7f68d2dff435b93b21020696655d6f1e51b7b5aef2265ff142ee679b25","cross_cats_sorted":["cs.LG"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-06-17T17:13:08Z","title_canon_sha256":"b53e69677dd2f001872252684ee15e15817327f405fe48d5d16c0bd6e3855296"},"schema_version":"1.0","source":{"id":"2406.11753","kind":"arxiv","version":3}},"canonical_sha256":"5255eec20a41fbdc92852906cefd3f8a2474a63b6c1873373b49fa59144cca55","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"5255eec20a41fbdc92852906cefd3f8a2474a63b6c1873373b49fa59144cca55","first_computed_at":"2026-07-05T11:13:10.708812Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:13:10.708812Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"BUmMS6oQrK8IlpXENzdS3JDcZM96HsL3EdBX3MCXoFoYF7Xz6WBe5gjBXv/E5ACMQwc7VTDvt0OWKimPPhyLBg==","signature_status":"signed_v1","signed_at":"2026-07-05T11:13:10.709332Z","signed_message":"canonical_sha256_bytes"},"source_id":"2406.11753","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:21ccb1577584a0cdbf95fc895abcfd5c78985c03a1b407be613eb502d3d18ab1","sha256:461fb1defe8ae2466ec77e749f93c61318927c1cf61c3ba9faa3900b0a775dc5"],"state_sha256":"dbe4871902b7a789844c57ad853e6293497cf023caf6875f1482b65616857a6c"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"FIVxS7Ccp/suY/fPxkvntZukpMOXgK3ekOCUODv1SMzmTmV/wRdnbLWf1TIt0n0BDP4USfXEnG1igu3cEYlfAg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-09T15:30:00.071654Z","bundle_sha256":"f2c911a3040fc556058c9d8d0169b7a0ea40a011b9abd5182a8a90ebe4a4d715"}}