{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:C7AAK6BY5S64WSHHLARQ63N4LS","short_pith_number":"pith:C7AAK6BY","schema_version":"1.0","canonical_sha256":"17c0057838ecbdcb48e758230f6dbc5c87bb84880bbc191a84c22c20a8f143bb","source":{"kind":"arxiv","id":"2605.25704","version":1},"attestation_state":"computed","paper":{"title":"PowLU: An Activation Function for Stable Pre-Training of LLMs","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Cunyin Peng, Jia Liu, Jun Zhou, Kunlong Chen, Peijie Jiang, Qian Zhao, Yuqi Feng, Zhiqiang Zhang","submitted_at":"2026-05-25T11:02:05Z","abstract_excerpt":"In contemporary large language models (LLMs), the swish-gated linear unit (SwiGLU) activation function is widely adopted to regulate the information flow and introduce non-linearity. For large positive inputs, SwiGLU approximates the quadratic function $x^2$, providing strong nonlinearity and expressive capacity. However, this property also causes numerical instability as the input or model scale increases, particularly in low-precision LLM training. The main reason is its approximate quadratic amplification, which enlarges the output range and exacerbates outliers. To address this issue, we p"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2605.25704","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2026-05-25T11:02:05Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"7654c9f34606c399dc842ab99d24c44f89f6a62aa8feddd0c0760d26058ce828","abstract_canon_sha256":"bae395fcf3e6267ba8519dcf761a1fd41a4cbe5289a6947b437a181ddae2969b"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-26T02:04:51.386430Z","signature_b64":"51XgP8nWHQKxqqoozvAzXrswOy8s3SIXU761sqGGyzXV9xJ7FCEN4zkhb47ACmHiBTiUdnxxyzHcLdhVv541Dg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"17c0057838ecbdcb48e758230f6dbc5c87bb84880bbc191a84c22c20a8f143bb","last_reissued_at":"2026-05-26T02:04:51.385929Z","signature_status":"signed_v1","first_computed_at":"2026-05-26T02:04:51.385929Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"PowLU: An Activation Function for Stable Pre-Training of LLMs","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Cunyin Peng, Jia Liu, Jun Zhou, Kunlong Chen, Peijie Jiang, Qian Zhao, Yuqi Feng, Zhiqiang Zhang","submitted_at":"2026-05-25T11:02:05Z","abstract_excerpt":"In contemporary large language models (LLMs), the swish-gated linear unit (SwiGLU) activation function is widely adopted to regulate the information flow and introduce non-linearity. For large positive inputs, SwiGLU approximates the quadratic function $x^2$, providing strong nonlinearity and expressive capacity. However, this property also causes numerical instability as the input or model scale increases, particularly in low-precision LLM training. The main reason is its approximate quadratic amplification, which enlarges the output range and exacerbates outliers. To address this issue, we p"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.25704","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/2605.25704/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"},"aliases":[{"alias_kind":"arxiv","alias_value":"2605.25704","created_at":"2026-05-26T02:04:51.385999+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.25704v1","created_at":"2026-05-26T02:04:51.385999+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.25704","created_at":"2026-05-26T02:04:51.385999+00:00"},{"alias_kind":"pith_short_12","alias_value":"C7AAK6BY5S64","created_at":"2026-05-26T02:04:51.385999+00:00"},{"alias_kind":"pith_short_16","alias_value":"C7AAK6BY5S64WSHH","created_at":"2026-05-26T02:04:51.385999+00:00"},{"alias_kind":"pith_short_8","alias_value":"C7AAK6BY","created_at":"2026-05-26T02:04:51.385999+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/C7AAK6BY5S64WSHHLARQ63N4LS","json":"https://pith.science/pith/C7AAK6BY5S64WSHHLARQ63N4LS.json","graph_json":"https://pith.science/api/pith-number/C7AAK6BY5S64WSHHLARQ63N4LS/graph.json","events_json":"https://pith.science/api/pith-number/C7AAK6BY5S64WSHHLARQ63N4LS/events.json","paper":"https://pith.science/paper/C7AAK6BY"},"agent_actions":{"view_html":"https://pith.science/pith/C7AAK6BY5S64WSHHLARQ63N4LS","download_json":"https://pith.science/pith/C7AAK6BY5S64WSHHLARQ63N4LS.json","view_paper":"https://pith.science/paper/C7AAK6BY","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.25704&json=true","fetch_graph":"https://pith.science/api/pith-number/C7AAK6BY5S64WSHHLARQ63N4LS/graph.json","fetch_events":"https://pith.science/api/pith-number/C7AAK6BY5S64WSHHLARQ63N4LS/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/C7AAK6BY5S64WSHHLARQ63N4LS/action/timestamp_anchor","attest_storage":"https://pith.science/pith/C7AAK6BY5S64WSHHLARQ63N4LS/action/storage_attestation","attest_author":"https://pith.science/pith/C7AAK6BY5S64WSHHLARQ63N4LS/action/author_attestation","sign_citation":"https://pith.science/pith/C7AAK6BY5S64WSHHLARQ63N4LS/action/citation_signature","submit_replication":"https://pith.science/pith/C7AAK6BY5S64WSHHLARQ63N4LS/action/replication_record"}},"created_at":"2026-05-26T02:04:51.385999+00:00","updated_at":"2026-05-26T02:04:51.385999+00:00"}