{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:TC6COLDCLB52AFS7LGF2QAIXHD","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":"c8c3fae2a5fcffb71bb4682697ac08fe566615f1f40760974cb93ff14140c049","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-05-29T17:11:28Z","title_canon_sha256":"7a3a0bf92294bf9160b0a52ee6525f5cc11be60352a9f9e61504fd94c3efd01f"},"schema_version":"1.0","source":{"id":"2405.19279","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2405.19279","created_at":"2026-07-05T09:32:09Z"},{"alias_kind":"arxiv_version","alias_value":"2405.19279v2","created_at":"2026-07-05T09:32:09Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2405.19279","created_at":"2026-07-05T09:32:09Z"},{"alias_kind":"pith_short_12","alias_value":"TC6COLDCLB52","created_at":"2026-07-05T09:32:09Z"},{"alias_kind":"pith_short_16","alias_value":"TC6COLDCLB52AFS7","created_at":"2026-07-05T09:32:09Z"},{"alias_kind":"pith_short_8","alias_value":"TC6COLDC","created_at":"2026-07-05T09:32:09Z"}],"graph_snapshots":[{"event_id":"sha256:c25cb45d465c3001b9be06f2dce0be7f8b858b7cd44fbd07ceb1db4829088ff8","target":"graph","created_at":"2026-07-05T09:32:09Z","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/2405.19279/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Outlier Features (OFs) are neurons whose activation magnitudes significantly exceed the average over a neural network's (NN) width. They are well known to emerge during standard transformer training and have the undesirable effect of hindering quantisation in afflicted models. Despite their practical importance, little is known behind why OFs emerge during training, nor how one can minimise them.\n  Our work focuses on the above questions, first identifying several quantitative metrics, such as the kurtosis over neuron activation norms, to measure OFs. With these metrics, we study how architect","authors_text":"Bobby He, Daniele Paliotta, Imanol Schlag, Lorenzo Noci, Thomas Hofmann","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-05-29T17:11:28Z","title":"Understanding and Minimising Outlier Features in Neural Network Training"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2405.19279","kind":"arxiv","version":2},"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:edabe3e320761886cca2244b4e6c41b865bf85c930025e6794c05c7f6620c917","target":"record","created_at":"2026-07-05T09:32:09Z","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":"c8c3fae2a5fcffb71bb4682697ac08fe566615f1f40760974cb93ff14140c049","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-05-29T17:11:28Z","title_canon_sha256":"7a3a0bf92294bf9160b0a52ee6525f5cc11be60352a9f9e61504fd94c3efd01f"},"schema_version":"1.0","source":{"id":"2405.19279","kind":"arxiv","version":2}},"canonical_sha256":"98bc272c62587ba0165f598ba8011738e60d735123a179216a8000c381b35897","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"98bc272c62587ba0165f598ba8011738e60d735123a179216a8000c381b35897","first_computed_at":"2026-07-05T09:32:09.609232Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T09:32:09.609232Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"2lBnOUjWmewv87r8+sd8CWdcrmle47DwwaUxtQnWAXNhapNSMXsaKDRh+/CAmuogVvp9uNNeh4cRY66uzLibCw==","signature_status":"signed_v1","signed_at":"2026-07-05T09:32:09.609763Z","signed_message":"canonical_sha256_bytes"},"source_id":"2405.19279","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:edabe3e320761886cca2244b4e6c41b865bf85c930025e6794c05c7f6620c917","sha256:c25cb45d465c3001b9be06f2dce0be7f8b858b7cd44fbd07ceb1db4829088ff8"],"state_sha256":"df59f897def724f3f20f78d638a652b358d468fdf06d8a5ea91beb3237796ecb"}