{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:4H6GSHNSJ2ZLFG72JKVQ466ULY","short_pith_number":"pith:4H6GSHNS","canonical_record":{"source":{"id":"2508.06016","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2025-08-08T05:04:28Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"740679fa50d89ff13057b007daec442017291ec1c34c66c7a4d40021dd352c53","abstract_canon_sha256":"da8536334e07b7721fdbedc263a213295b23edbf1921f53f7cab06857630982c"},"schema_version":"1.0"},"canonical_sha256":"e1fc691db24eb2b29bfa4aab0e7bd45e2cbdc650a39afe540a0b115a18f83470","source":{"kind":"arxiv","id":"2508.06016","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2508.06016","created_at":"2026-07-05T11:50:47Z"},{"alias_kind":"arxiv_version","alias_value":"2508.06016v1","created_at":"2026-07-05T11:50:47Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2508.06016","created_at":"2026-07-05T11:50:47Z"},{"alias_kind":"pith_short_12","alias_value":"4H6GSHNSJ2ZL","created_at":"2026-07-05T11:50:47Z"},{"alias_kind":"pith_short_16","alias_value":"4H6GSHNSJ2ZLFG72","created_at":"2026-07-05T11:50:47Z"},{"alias_kind":"pith_short_8","alias_value":"4H6GSHNS","created_at":"2026-07-05T11:50:47Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:4H6GSHNSJ2ZLFG72JKVQ466ULY","target":"record","payload":{"canonical_record":{"source":{"id":"2508.06016","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2025-08-08T05:04:28Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"740679fa50d89ff13057b007daec442017291ec1c34c66c7a4d40021dd352c53","abstract_canon_sha256":"da8536334e07b7721fdbedc263a213295b23edbf1921f53f7cab06857630982c"},"schema_version":"1.0"},"canonical_sha256":"e1fc691db24eb2b29bfa4aab0e7bd45e2cbdc650a39afe540a0b115a18f83470","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:50:47.445952Z","signature_b64":"8osVP8iUcYDS4OA92lnRYCiKqQS7r5qN3o8JishyEClASPr5mzrXY6WG9VuJQeeRj8zMxXlwSrvrJJQ/sT2HCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e1fc691db24eb2b29bfa4aab0e7bd45e2cbdc650a39afe540a0b115a18f83470","last_reissued_at":"2026-07-05T11:50:47.445427Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:50:47.445427Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2508.06016","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":"2026-07-05T11:50:47Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"+eOD7v5hjbP0O0D1dWCdqnB+uWQW0FL8KxxXF7mqQ2EQ0TEiTEeFPU14CoHfy6VvfRbHMBmOwBxovw7PR6UKAA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-07T03:58:11.197125Z"},"content_sha256":"23715fb8b428877a8f5297a0ad78e41a28c233ce6f4f9cf1e1e71d88f2120ac4","schema_version":"1.0","event_id":"sha256:23715fb8b428877a8f5297a0ad78e41a28c233ce6f4f9cf1e1e71d88f2120ac4"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:4H6GSHNSJ2ZLFG72JKVQ466ULY","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Crisp Attention: Regularizing Transformers via Structured Sparsity","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Sagar Gandhi, Vishal Gandhi","submitted_at":"2025-08-08T05:04:28Z","abstract_excerpt":"The quadratic computational cost of the self-attention mechanism is a primary challenge in scaling Transformer models. While attention sparsity is widely studied as a technique to improve computational efficiency, it is almost universally assumed to come at the cost of model accuracy. In this paper, we report a surprising counter-example to this common wisdom. By introducing structured, post-hoc sparsity to the attention mechanism of a DistilBERT model during fine-tuning on the SST-2 sentiment analysis task, we find that model accuracy improves significantly. Our model with 80\\% attention spar"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2508.06016","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/2508.06016/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":"2026-07-05T11:50:47Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"CdLDmP9qJkv9m0EPv6RfoyWOtDhHI5CLWIRf9Oig8HHQLtktNkkTK7+DjwsAaWGw/7NquzrW7tyE5eyiLr2QDw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-07T03:58:11.197624Z"},"content_sha256":"9957a69cbd2c5c536b0a7f4f658cf9c1883fb0a8b68708e4b83cb866f1a4ded5","schema_version":"1.0","event_id":"sha256:9957a69cbd2c5c536b0a7f4f658cf9c1883fb0a8b68708e4b83cb866f1a4ded5"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/4H6GSHNSJ2ZLFG72JKVQ466ULY/bundle.json","state_url":"https://pith.science/pith/4H6GSHNSJ2ZLFG72JKVQ466ULY/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/4H6GSHNSJ2ZLFG72JKVQ466ULY/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":"2026-08-07T03:58:11Z","links":{"resolver":"https://pith.science/pith/4H6GSHNSJ2ZLFG72JKVQ466ULY","bundle":"https://pith.science/pith/4H6GSHNSJ2ZLFG72JKVQ466ULY/bundle.json","state":"https://pith.science/pith/4H6GSHNSJ2ZLFG72JKVQ466ULY/state.json","well_known_bundle":"https://pith.science/.well-known/pith/4H6GSHNSJ2ZLFG72JKVQ466ULY/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:4H6GSHNSJ2ZLFG72JKVQ466ULY","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":"da8536334e07b7721fdbedc263a213295b23edbf1921f53f7cab06857630982c","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2025-08-08T05:04:28Z","title_canon_sha256":"740679fa50d89ff13057b007daec442017291ec1c34c66c7a4d40021dd352c53"},"schema_version":"1.0","source":{"id":"2508.06016","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2508.06016","created_at":"2026-07-05T11:50:47Z"},{"alias_kind":"arxiv_version","alias_value":"2508.06016v1","created_at":"2026-07-05T11:50:47Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2508.06016","created_at":"2026-07-05T11:50:47Z"},{"alias_kind":"pith_short_12","alias_value":"4H6GSHNSJ2ZL","created_at":"2026-07-05T11:50:47Z"},{"alias_kind":"pith_short_16","alias_value":"4H6GSHNSJ2ZLFG72","created_at":"2026-07-05T11:50:47Z"},{"alias_kind":"pith_short_8","alias_value":"4H6GSHNS","created_at":"2026-07-05T11:50:47Z"}],"graph_snapshots":[{"event_id":"sha256:9957a69cbd2c5c536b0a7f4f658cf9c1883fb0a8b68708e4b83cb866f1a4ded5","target":"graph","created_at":"2026-07-05T11:50:47Z","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/2508.06016/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"The quadratic computational cost of the self-attention mechanism is a primary challenge in scaling Transformer models. While attention sparsity is widely studied as a technique to improve computational efficiency, it is almost universally assumed to come at the cost of model accuracy. In this paper, we report a surprising counter-example to this common wisdom. By introducing structured, post-hoc sparsity to the attention mechanism of a DistilBERT model during fine-tuning on the SST-2 sentiment analysis task, we find that model accuracy improves significantly. Our model with 80\\% attention spar","authors_text":"Sagar Gandhi, Vishal Gandhi","cross_cats":["cs.AI"],"headline":"","license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2025-08-08T05:04:28Z","title":"Crisp Attention: Regularizing Transformers via Structured Sparsity"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2508.06016","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":"sha256:23715fb8b428877a8f5297a0ad78e41a28c233ce6f4f9cf1e1e71d88f2120ac4","target":"record","created_at":"2026-07-05T11:50:47Z","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":"da8536334e07b7721fdbedc263a213295b23edbf1921f53f7cab06857630982c","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2025-08-08T05:04:28Z","title_canon_sha256":"740679fa50d89ff13057b007daec442017291ec1c34c66c7a4d40021dd352c53"},"schema_version":"1.0","source":{"id":"2508.06016","kind":"arxiv","version":1}},"canonical_sha256":"e1fc691db24eb2b29bfa4aab0e7bd45e2cbdc650a39afe540a0b115a18f83470","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"e1fc691db24eb2b29bfa4aab0e7bd45e2cbdc650a39afe540a0b115a18f83470","first_computed_at":"2026-07-05T11:50:47.445427Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:50:47.445427Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"8osVP8iUcYDS4OA92lnRYCiKqQS7r5qN3o8JishyEClASPr5mzrXY6WG9VuJQeeRj8zMxXlwSrvrJJQ/sT2HCA==","signature_status":"signed_v1","signed_at":"2026-07-05T11:50:47.445952Z","signed_message":"canonical_sha256_bytes"},"source_id":"2508.06016","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:23715fb8b428877a8f5297a0ad78e41a28c233ce6f4f9cf1e1e71d88f2120ac4","sha256:9957a69cbd2c5c536b0a7f4f658cf9c1883fb0a8b68708e4b83cb866f1a4ded5"],"state_sha256":"0cb0f189ef274daf2948109776be00b07fdd9859dd0346f495bb48374d1cfa41"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"SzO2OXTAi9567mXYYPNnkuUW4DLkbomHpLxFp2tWDNvibLscm06hvj2uA5ClcRBpLcE8yZ+j0MtDYSv7hawFAQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-07T03:58:11.202153Z","bundle_sha256":"6070d1fdd2962dcdd8d8d0de23d834ffed6a967c85ba79dc48d6e4d92c05dfec"}}