{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:KN4PYLFXIGK57LCX22YPNBZLC6","short_pith_number":"pith:KN4PYLFX","canonical_record":{"source":{"id":"2404.09695","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-04-15T11:53:22Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"f09bba9b3b92a36816bae8560eb511f36ad7f5e043d9560e650f8ff487dc086b","abstract_canon_sha256":"16edfad10a7d4bb7ab6f4dcd92e952b84f019e545ee9a53c2724f7bda14178bd"},"schema_version":"1.0"},"canonical_sha256":"5378fc2cb74195dfac57d6b0f6872b178d74ec1196bff9a699a6543575770238","source":{"kind":"arxiv","id":"2404.09695","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2404.09695","created_at":"2026-07-05T08:08:10Z"},{"alias_kind":"arxiv_version","alias_value":"2404.09695v1","created_at":"2026-07-05T08:08:10Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2404.09695","created_at":"2026-07-05T08:08:10Z"},{"alias_kind":"pith_short_12","alias_value":"KN4PYLFXIGK5","created_at":"2026-07-05T08:08:10Z"},{"alias_kind":"pith_short_16","alias_value":"KN4PYLFXIGK57LCX","created_at":"2026-07-05T08:08:10Z"},{"alias_kind":"pith_short_8","alias_value":"KN4PYLFX","created_at":"2026-07-05T08:08:10Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:KN4PYLFXIGK57LCX22YPNBZLC6","target":"record","payload":{"canonical_record":{"source":{"id":"2404.09695","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-04-15T11:53:22Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"f09bba9b3b92a36816bae8560eb511f36ad7f5e043d9560e650f8ff487dc086b","abstract_canon_sha256":"16edfad10a7d4bb7ab6f4dcd92e952b84f019e545ee9a53c2724f7bda14178bd"},"schema_version":"1.0"},"canonical_sha256":"5378fc2cb74195dfac57d6b0f6872b178d74ec1196bff9a699a6543575770238","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:08:10.476380Z","signature_b64":"Y+LNdiNqn8jkIWd6GTsG2HfaLy9wrDbLslPmgyQ8TB0dqONVOvfpJNUvWD0Pc/+z0q6kWzJbmZerbFIfyBGLAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"5378fc2cb74195dfac57d6b0f6872b178d74ec1196bff9a699a6543575770238","last_reissued_at":"2026-07-05T08:08:10.475822Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:08:10.475822Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2404.09695","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":"2026-07-05T08:08:10Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"atXOx8uLZbbaOjbSBUIAmns3J0E043+JJQE7Kf2dimWgV9Iba+q8ujRZdYF2mG4+J/xdiTgbQgVvDY+1LWniBw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-04T15:05:51.711954Z"},"content_sha256":"3273d9c1fffd55ec072c9a1eb29a97aed3e1e051b1b6eed4637ef69d634b1545","schema_version":"1.0","event_id":"sha256:3273d9c1fffd55ec072c9a1eb29a97aed3e1e051b1b6eed4637ef69d634b1545"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:KN4PYLFXIGK57LCX22YPNBZLC6","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"LoRAP: Transformer Sub-Layers Deserve Differentiated Structured Compression for Large Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.LG","authors_text":"Guangyan Li, Wensheng Zhang, Yongqiang Tang","submitted_at":"2024-04-15T11:53:22Z","abstract_excerpt":"Large language models (LLMs) show excellent performance in difficult tasks, but they often require massive memories and computational resources. How to reduce the parameter scale of LLMs has become research hotspots. In this study, we make an important observation that the multi-head self-attention (MHA) sub-layer of Transformer exhibits noticeable low-rank structure, while the feed-forward network (FFN) sub-layer does not. With this regard, we design a mixed compression model, which organically combines Low-Rank matrix approximation And structured Pruning (LoRAP). For the MHA sub-layer, we pr"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2404.09695","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/2404.09695/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":"2026-07-05T08:08:10Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"DJ/vo8A3AIyrnuWvn+jTnHZoG4Bp03ABlq6+IO/Gjlr0clWWUi8HsZf8iEsyFI2bprV/4hWdzWSe6EIhmBrtCg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-04T15:05:51.712795Z"},"content_sha256":"280e62d7ac106f8b405ba26e4f7df685a795299308d595926f8cca78ebfcebff","schema_version":"1.0","event_id":"sha256:280e62d7ac106f8b405ba26e4f7df685a795299308d595926f8cca78ebfcebff"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/KN4PYLFXIGK57LCX22YPNBZLC6/bundle.json","state_url":"https://pith.science/pith/KN4PYLFXIGK57LCX22YPNBZLC6/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/KN4PYLFXIGK57LCX22YPNBZLC6/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":"2026-08-04T15:05:51Z","links":{"resolver":"https://pith.science/pith/KN4PYLFXIGK57LCX22YPNBZLC6","bundle":"https://pith.science/pith/KN4PYLFXIGK57LCX22YPNBZLC6/bundle.json","state":"https://pith.science/pith/KN4PYLFXIGK57LCX22YPNBZLC6/state.json","well_known_bundle":"https://pith.science/.well-known/pith/KN4PYLFXIGK57LCX22YPNBZLC6/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:KN4PYLFXIGK57LCX22YPNBZLC6","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":"16edfad10a7d4bb7ab6f4dcd92e952b84f019e545ee9a53c2724f7bda14178bd","cross_cats_sorted":["cs.AI","cs.CL"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-04-15T11:53:22Z","title_canon_sha256":"f09bba9b3b92a36816bae8560eb511f36ad7f5e043d9560e650f8ff487dc086b"},"schema_version":"1.0","source":{"id":"2404.09695","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2404.09695","created_at":"2026-07-05T08:08:10Z"},{"alias_kind":"arxiv_version","alias_value":"2404.09695v1","created_at":"2026-07-05T08:08:10Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2404.09695","created_at":"2026-07-05T08:08:10Z"},{"alias_kind":"pith_short_12","alias_value":"KN4PYLFXIGK5","created_at":"2026-07-05T08:08:10Z"},{"alias_kind":"pith_short_16","alias_value":"KN4PYLFXIGK57LCX","created_at":"2026-07-05T08:08:10Z"},{"alias_kind":"pith_short_8","alias_value":"KN4PYLFX","created_at":"2026-07-05T08:08:10Z"}],"graph_snapshots":[{"event_id":"sha256:280e62d7ac106f8b405ba26e4f7df685a795299308d595926f8cca78ebfcebff","target":"graph","created_at":"2026-07-05T08:08:10Z","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/2404.09695/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Large language models (LLMs) show excellent performance in difficult tasks, but they often require massive memories and computational resources. How to reduce the parameter scale of LLMs has become research hotspots. In this study, we make an important observation that the multi-head self-attention (MHA) sub-layer of Transformer exhibits noticeable low-rank structure, while the feed-forward network (FFN) sub-layer does not. With this regard, we design a mixed compression model, which organically combines Low-Rank matrix approximation And structured Pruning (LoRAP). For the MHA sub-layer, we pr","authors_text":"Guangyan Li, Wensheng Zhang, Yongqiang Tang","cross_cats":["cs.AI","cs.CL"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-04-15T11:53:22Z","title":"LoRAP: Transformer Sub-Layers Deserve Differentiated Structured Compression for Large Language Models"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2404.09695","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":"sha256:3273d9c1fffd55ec072c9a1eb29a97aed3e1e051b1b6eed4637ef69d634b1545","target":"record","created_at":"2026-07-05T08:08:10Z","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":"16edfad10a7d4bb7ab6f4dcd92e952b84f019e545ee9a53c2724f7bda14178bd","cross_cats_sorted":["cs.AI","cs.CL"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-04-15T11:53:22Z","title_canon_sha256":"f09bba9b3b92a36816bae8560eb511f36ad7f5e043d9560e650f8ff487dc086b"},"schema_version":"1.0","source":{"id":"2404.09695","kind":"arxiv","version":1}},"canonical_sha256":"5378fc2cb74195dfac57d6b0f6872b178d74ec1196bff9a699a6543575770238","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"5378fc2cb74195dfac57d6b0f6872b178d74ec1196bff9a699a6543575770238","first_computed_at":"2026-07-05T08:08:10.475822Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T08:08:10.475822Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"Y+LNdiNqn8jkIWd6GTsG2HfaLy9wrDbLslPmgyQ8TB0dqONVOvfpJNUvWD0Pc/+z0q6kWzJbmZerbFIfyBGLAw==","signature_status":"signed_v1","signed_at":"2026-07-05T08:08:10.476380Z","signed_message":"canonical_sha256_bytes"},"source_id":"2404.09695","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:3273d9c1fffd55ec072c9a1eb29a97aed3e1e051b1b6eed4637ef69d634b1545","sha256:280e62d7ac106f8b405ba26e4f7df685a795299308d595926f8cca78ebfcebff"],"state_sha256":"3ffb9207a22696344de49268f1a839fa51c8f67eaf03dbce30e03e7be0fde8c4"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"yoYwRKBI1vS3wVqoNzd7JRyfaJ2f1MlQ7caKHUIT5U9fTSpGcyEbtnQBUW1I5XeweNOH04V++YhS77IPibaFAA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-04T15:05:51.718749Z","bundle_sha256":"2405e9393b94ba55f324241b7d75e6dd11f22229fa10d6329af6ae557eae72b5"}}