{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:L7NYTNSTZJXPIILKGOVVKEXZ54","short_pith_number":"pith:L7NYTNST","canonical_record":{"source":{"id":"2605.06152","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-07T12:45:21Z","cross_cats_sorted":["cs.CL","math.OC","stat.ML"],"title_canon_sha256":"b286a045e76c4eba4a7313944c8d26b736e3eb958e213f3b7b49d9e25c003486","abstract_canon_sha256":"30d360ecf2d6eab91e508e8d66faefea61f8ca43dab8def0f9c1ed9e16f7486a"},"schema_version":"1.0"},"canonical_sha256":"5fdb89b653ca6ef4216a33ab5512f9ef29b3b23ece549439da00d82227a108e7","source":{"kind":"arxiv","id":"2605.06152","version":3},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.06152","created_at":"2026-05-27T01:05:56Z"},{"alias_kind":"arxiv_version","alias_value":"2605.06152v3","created_at":"2026-05-27T01:05:56Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.06152","created_at":"2026-05-27T01:05:56Z"},{"alias_kind":"pith_short_12","alias_value":"L7NYTNSTZJXP","created_at":"2026-05-27T01:05:56Z"},{"alias_kind":"pith_short_16","alias_value":"L7NYTNSTZJXPIILK","created_at":"2026-05-27T01:05:56Z"},{"alias_kind":"pith_short_8","alias_value":"L7NYTNST","created_at":"2026-05-27T01:05:56Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:L7NYTNSTZJXPIILKGOVVKEXZ54","target":"record","payload":{"canonical_record":{"source":{"id":"2605.06152","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-07T12:45:21Z","cross_cats_sorted":["cs.CL","math.OC","stat.ML"],"title_canon_sha256":"b286a045e76c4eba4a7313944c8d26b736e3eb958e213f3b7b49d9e25c003486","abstract_canon_sha256":"30d360ecf2d6eab91e508e8d66faefea61f8ca43dab8def0f9c1ed9e16f7486a"},"schema_version":"1.0"},"canonical_sha256":"5fdb89b653ca6ef4216a33ab5512f9ef29b3b23ece549439da00d82227a108e7","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-27T01:05:56.308281Z","signature_b64":"YNphSdfvrdNPvNgVizQj3z+bV3pAVjwAOnKQUtl8pzW/Bwf39ZL/UVOzKKMUKbi4HGeSTylPWd0bpgZJTaAjAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"5fdb89b653ca6ef4216a33ab5512f9ef29b3b23ece549439da00d82227a108e7","last_reissued_at":"2026-05-27T01:05:56.307519Z","signature_status":"signed_v1","first_computed_at":"2026-05-27T01:05:56.307519Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2605.06152","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":"2026-05-27T01:05:56Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"RQytCdlPS3OAXsXeL6+Y8h1791wLKI4E8V4WtFNpvl9+K8z1nDfYWAFGUf/HiEz7hwAMxtGGCD1zY2MpMg6zCA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-08T21:54:09.120508Z"},"content_sha256":"1b304fec4b696a2c08906e196f64fa4e16548588cc3df29926cec550faca6cb1","schema_version":"1.0","event_id":"sha256:1b304fec4b696a2c08906e196f64fa4e16548588cc3df29926cec550faca6cb1"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:L7NYTNSTZJXPIILKGOVVKEXZ54","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Grokking or Glitching? How Low-Precision Drives Slingshot Loss Spikes","license":"http://creativecommons.org/licenses/by/4.0/","headline":"Floating-point precision limits trigger slingshot loss spikes by creating numerical feature inflation in neural network training.","cross_cats":["cs.CL","math.OC","stat.ML"],"primary_cat":"cs.LG","authors_text":"Jianjun Cao, Liu Hanqing, Yuanze Li, Zijian Zhou","submitted_at":"2026-05-07T12:45:21Z","abstract_excerpt":"Deep neural networks exhibit periodic loss spikes during unregularized long-term training, a phenomenon known as the \"Slingshot Mechanism.\" Existing work usually attributes this to intrinsic optimization dynamics, but its triggering mechanism remains unclear. This paper proves that this phenomenon is a result of floating-point arithmetic precision limits. As training enters a high-confidence stage, the difference between the correct-class logit and the other logits may exceed the absorption-error threshold. Then during backpropagation, the gradient of the correct class is rounded exactly to ze"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"This paper proves that this phenomenon is a result of floating-point arithmetic precision limits. ... We prove that this drift forms a positive feedback loop with the feature, causing the global classifier mean and the global feature mean to grow exponentially.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"The assumption that, once the logit difference exceeds the absorption-error threshold, the gradient of the correct class is rounded exactly to zero during backpropagation while incorrect-class gradients remain nonzero, and that this imbalance necessarily creates an exponential positive feedback loop with the features.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"Slingshot loss spikes arise from floating-point precision limits that round correct-class gradients to zero, breaking zero-sum constraints and driving exponential parameter growth through numerical feature inflation.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Floating-point precision limits trigger slingshot loss spikes by creating numerical feature inflation in neural network training.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"4403ecc533f8b8e57dc16cf4bc478615390e44a5f2ff1722747c74a475579148"},"source":{"id":"2605.06152","kind":"arxiv","version":3},"verdict":{"id":"73fe5ca6-aeea-4410-b502-b00c8227078a","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-13T07:06:05.189778Z","strongest_claim":"This paper proves that this phenomenon is a result of floating-point arithmetic precision limits. ... We prove that this drift forms a positive feedback loop with the feature, causing the global classifier mean and the global feature mean to grow exponentially.","one_line_summary":"Slingshot loss spikes arise from floating-point precision limits that round correct-class gradients to zero, breaking zero-sum constraints and driving exponential parameter growth through numerical feature inflation.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"The assumption that, once the logit difference exceeds the absorption-error threshold, the gradient of the correct class is rounded exactly to zero during backpropagation while incorrect-class gradients remain nonzero, and that this imbalance necessarily creates an exponential positive feedback loop with the features.","pith_extraction_headline":"Floating-point precision limits trigger slingshot loss spikes by creating numerical feature inflation in neural network training."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2605.06152/integrity.json","findings":[],"available":true,"detectors_run":[{"name":"claim_evidence","ran_at":"2026-05-20T13:02:04.285171Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"ai_meta_artifact","ran_at":"2026-05-20T08:36:17.189756Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_title_agreement","ran_at":"2026-05-19T19:01:19.348912Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_compliance","ran_at":"2026-05-19T12:55:23.639732Z","status":"completed","version":"1.0.0","findings_count":0}],"snapshot_sha256":"5704dbfe57f64c954f96da93292ab57ce4e9d70270532fa950c2c4fa2712e08a"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":2,"snapshot_sha256":"b9b5ab6c7044a478fdc62f310b3c9f9d4878575de948cce219a613dc5ec39778"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"verdict_id":"73fe5ca6-aeea-4410-b502-b00c8227078a"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-27T01:05:56Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"EPjTg5fsLO4wqfGABvyY7jVuhki3jpfltvON2T+OfEZ/jemBA4ajePuRWhYkFGPNIwlOhPpBpShZgnHUlSQdBw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-08T21:54:09.121013Z"},"content_sha256":"69d37531958054660cff35ae3ad2256ee356afe0d16e5dddc193c5efe34253ab","schema_version":"1.0","event_id":"sha256:69d37531958054660cff35ae3ad2256ee356afe0d16e5dddc193c5efe34253ab"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/L7NYTNSTZJXPIILKGOVVKEXZ54/bundle.json","state_url":"https://pith.science/pith/L7NYTNSTZJXPIILKGOVVKEXZ54/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/L7NYTNSTZJXPIILKGOVVKEXZ54/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":"2026-06-08T21:54:09Z","links":{"resolver":"https://pith.science/pith/L7NYTNSTZJXPIILKGOVVKEXZ54","bundle":"https://pith.science/pith/L7NYTNSTZJXPIILKGOVVKEXZ54/bundle.json","state":"https://pith.science/pith/L7NYTNSTZJXPIILKGOVVKEXZ54/state.json","well_known_bundle":"https://pith.science/.well-known/pith/L7NYTNSTZJXPIILKGOVVKEXZ54/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:L7NYTNSTZJXPIILKGOVVKEXZ54","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":"30d360ecf2d6eab91e508e8d66faefea61f8ca43dab8def0f9c1ed9e16f7486a","cross_cats_sorted":["cs.CL","math.OC","stat.ML"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-07T12:45:21Z","title_canon_sha256":"b286a045e76c4eba4a7313944c8d26b736e3eb958e213f3b7b49d9e25c003486"},"schema_version":"1.0","source":{"id":"2605.06152","kind":"arxiv","version":3}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.06152","created_at":"2026-05-27T01:05:56Z"},{"alias_kind":"arxiv_version","alias_value":"2605.06152v3","created_at":"2026-05-27T01:05:56Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.06152","created_at":"2026-05-27T01:05:56Z"},{"alias_kind":"pith_short_12","alias_value":"L7NYTNSTZJXP","created_at":"2026-05-27T01:05:56Z"},{"alias_kind":"pith_short_16","alias_value":"L7NYTNSTZJXPIILK","created_at":"2026-05-27T01:05:56Z"},{"alias_kind":"pith_short_8","alias_value":"L7NYTNST","created_at":"2026-05-27T01:05:56Z"}],"graph_snapshots":[{"event_id":"sha256:69d37531958054660cff35ae3ad2256ee356afe0d16e5dddc193c5efe34253ab","target":"graph","created_at":"2026-05-27T01:05:56Z","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":"This paper proves that this phenomenon is a result of floating-point arithmetic precision limits. ... We prove that this drift forms a positive feedback loop with the feature, causing the global classifier mean and the global feature mean to grow exponentially."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"The assumption that, once the logit difference exceeds the absorption-error threshold, the gradient of the correct class is rounded exactly to zero during backpropagation while incorrect-class gradients remain nonzero, and that this imbalance necessarily creates an exponential positive feedback loop with the features."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"Slingshot loss spikes arise from floating-point precision limits that round correct-class gradients to zero, breaking zero-sum constraints and driving exponential parameter growth through numerical feature inflation."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"Floating-point precision limits trigger slingshot loss spikes by creating numerical feature inflation in neural network training."}],"snapshot_sha256":"4403ecc533f8b8e57dc16cf4bc478615390e44a5f2ff1722747c74a475579148"},"formal_canon":{"evidence_count":2,"snapshot_sha256":"b9b5ab6c7044a478fdc62f310b3c9f9d4878575de948cce219a613dc5ec39778"},"integrity":{"available":true,"clean":true,"detectors_run":[{"findings_count":0,"name":"claim_evidence","ran_at":"2026-05-20T13:02:04.285171Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"ai_meta_artifact","ran_at":"2026-05-20T08:36:17.189756Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"doi_title_agreement","ran_at":"2026-05-19T19:01:19.348912Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"doi_compliance","ran_at":"2026-05-19T12:55:23.639732Z","status":"completed","version":"1.0.0"}],"endpoint":"/pith/2605.06152/integrity.json","findings":[],"snapshot_sha256":"5704dbfe57f64c954f96da93292ab57ce4e9d70270532fa950c2c4fa2712e08a","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Deep neural networks exhibit periodic loss spikes during unregularized long-term training, a phenomenon known as the \"Slingshot Mechanism.\" Existing work usually attributes this to intrinsic optimization dynamics, but its triggering mechanism remains unclear. This paper proves that this phenomenon is a result of floating-point arithmetic precision limits. As training enters a high-confidence stage, the difference between the correct-class logit and the other logits may exceed the absorption-error threshold. Then during backpropagation, the gradient of the correct class is rounded exactly to ze","authors_text":"Jianjun Cao, Liu Hanqing, Yuanze Li, Zijian Zhou","cross_cats":["cs.CL","math.OC","stat.ML"],"headline":"Floating-point precision limits trigger slingshot loss spikes by creating numerical feature inflation in neural network training.","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-07T12:45:21Z","title":"Grokking or Glitching? How Low-Precision Drives Slingshot Loss Spikes"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.06152","kind":"arxiv","version":3},"verdict":{"created_at":"2026-05-13T07:06:05.189778Z","id":"73fe5ca6-aeea-4410-b502-b00c8227078a","model_set":{"reader":"grok-4.3"},"one_line_summary":"Slingshot loss spikes arise from floating-point precision limits that round correct-class gradients to zero, breaking zero-sum constraints and driving exponential parameter growth through numerical feature inflation.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"Floating-point precision limits trigger slingshot loss spikes by creating numerical feature inflation in neural network training.","strongest_claim":"This paper proves that this phenomenon is a result of floating-point arithmetic precision limits. ... We prove that this drift forms a positive feedback loop with the feature, causing the global classifier mean and the global feature mean to grow exponentially.","weakest_assumption":"The assumption that, once the logit difference exceeds the absorption-error threshold, the gradient of the correct class is rounded exactly to zero during backpropagation while incorrect-class gradients remain nonzero, and that this imbalance necessarily creates an exponential positive feedback loop with the features."}},"verdict_id":"73fe5ca6-aeea-4410-b502-b00c8227078a"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:1b304fec4b696a2c08906e196f64fa4e16548588cc3df29926cec550faca6cb1","target":"record","created_at":"2026-05-27T01:05:56Z","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":"30d360ecf2d6eab91e508e8d66faefea61f8ca43dab8def0f9c1ed9e16f7486a","cross_cats_sorted":["cs.CL","math.OC","stat.ML"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-07T12:45:21Z","title_canon_sha256":"b286a045e76c4eba4a7313944c8d26b736e3eb958e213f3b7b49d9e25c003486"},"schema_version":"1.0","source":{"id":"2605.06152","kind":"arxiv","version":3}},"canonical_sha256":"5fdb89b653ca6ef4216a33ab5512f9ef29b3b23ece549439da00d82227a108e7","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"5fdb89b653ca6ef4216a33ab5512f9ef29b3b23ece549439da00d82227a108e7","first_computed_at":"2026-05-27T01:05:56.307519Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-27T01:05:56.307519Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"YNphSdfvrdNPvNgVizQj3z+bV3pAVjwAOnKQUtl8pzW/Bwf39ZL/UVOzKKMUKbi4HGeSTylPWd0bpgZJTaAjAg==","signature_status":"signed_v1","signed_at":"2026-05-27T01:05:56.308281Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.06152","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:1b304fec4b696a2c08906e196f64fa4e16548588cc3df29926cec550faca6cb1","sha256:69d37531958054660cff35ae3ad2256ee356afe0d16e5dddc193c5efe34253ab"],"state_sha256":"c9cdd1b87eab5722fc8fbb8322aaa33d9bd7ea62c19f983276f69f02b338b092"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"ESzMf2vFRYszVC5mOSvFz5EoDbnR8CLEoATFCAg2OaIdWwLlrKchRwraojW3fkm6Lu8A+ooHsWmzBSSc9h1ADA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-06-08T21:54:09.123292Z","bundle_sha256":"b625991a875e3ee653b1b2b4be833b9d2fa822a6ba5bd2dafa5e6d596dc36c30"}}