{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:PRUZSKIVAJWSJDS7X7RAOQDTBB","short_pith_number":"pith:PRUZSKIV","canonical_record":{"source":{"id":"2407.03651","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-07-04T05:46:20Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"a98d1f2e698caad6bacb002ff139954233293375d2ad5bbfa0c8508b8f29612d","abstract_canon_sha256":"97f220551d659f2e6279e4ccdbe7a3cdf4fa6c405ccbb2445683c8b75ce8b253"},"schema_version":"1.0"},"canonical_sha256":"7c69992915026d248e5fbfe207407308427bd53678499f67a79c65b047ed4538","source":{"kind":"arxiv","id":"2407.03651","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2407.03651","created_at":"2026-07-05T08:43:29Z"},{"alias_kind":"arxiv_version","alias_value":"2407.03651v2","created_at":"2026-07-05T08:43:29Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2407.03651","created_at":"2026-07-05T08:43:29Z"},{"alias_kind":"pith_short_12","alias_value":"PRUZSKIVAJWS","created_at":"2026-07-05T08:43:29Z"},{"alias_kind":"pith_short_16","alias_value":"PRUZSKIVAJWSJDS7","created_at":"2026-07-05T08:43:29Z"},{"alias_kind":"pith_short_8","alias_value":"PRUZSKIV","created_at":"2026-07-05T08:43:29Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:PRUZSKIVAJWSJDS7X7RAOQDTBB","target":"record","payload":{"canonical_record":{"source":{"id":"2407.03651","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-07-04T05:46:20Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"a98d1f2e698caad6bacb002ff139954233293375d2ad5bbfa0c8508b8f29612d","abstract_canon_sha256":"97f220551d659f2e6279e4ccdbe7a3cdf4fa6c405ccbb2445683c8b75ce8b253"},"schema_version":"1.0"},"canonical_sha256":"7c69992915026d248e5fbfe207407308427bd53678499f67a79c65b047ed4538","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:43:29.908511Z","signature_b64":"2PZDoFVagY3fA1xevXj8/s1CV9rsMWSr+TPCQA8i/my75B8oh0g9ALUJROaDlf/AbaHm3HL//vL7w9JtoxdiAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"7c69992915026d248e5fbfe207407308427bd53678499f67a79c65b047ed4538","last_reissued_at":"2026-07-05T08:43:29.908041Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:43:29.908041Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2407.03651","source_version":2,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T08:43:29Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"RRzRW/nhoD+Jl5/D8kuTM4wGk/MI/pDSzszQsPvXRx/VG+REACCMIJrViCgFt/WuWWptoSv74mDWZzPpcqsTDA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-07T18:22:58.761192Z"},"content_sha256":"a8306c8fb348daad45b59bda1e8b28178da34aa44a45e9213e0a6d9ea05840f1","schema_version":"1.0","event_id":"sha256:a8306c8fb348daad45b59bda1e8b28178da34aa44a45e9213e0a6d9ea05840f1"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:PRUZSKIVAJWSJDS7X7RAOQDTBB","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Evaluating Language Model Context Windows: A \"Working Memory\" Test and Inference-time Correction","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Amanda Dsouza, Changho Shin, Christopher Glaze, Frederic Sala","submitted_at":"2024-07-04T05:46:20Z","abstract_excerpt":"Large language models are prominently used in real-world applications, often tasked with reasoning over large volumes of documents. An exciting development in this space is models boasting extended context capabilities, with some accommodating over 2 million tokens. Such long context model capabilities remain uncertain in production systems, motivating the need to benchmark their performance on real world use cases. We address this challenge by proposing SWiM, an evaluation framework that addresses the limitations of standard tests. Testing the framework on eight long context models, we find t"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2407.03651","kind":"arxiv","version":2},"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/2407.03651/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":"2026-07-05T08:43:29Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"OaoyUyOc5jF3Sv5yC9DKPdQfzQ8XrFTnrqGyilRa8K2C4OxpwrPm7ul0/+gPB6X0NqW9xnJrNF+sZBzeIDj8Ag==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-07T18:22:58.761804Z"},"content_sha256":"5000809b80d84a3f27d8cf0a75515dbdca0971d95496e151e5a7e1537a36ff86","schema_version":"1.0","event_id":"sha256:5000809b80d84a3f27d8cf0a75515dbdca0971d95496e151e5a7e1537a36ff86"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/PRUZSKIVAJWSJDS7X7RAOQDTBB/bundle.json","state_url":"https://pith.science/pith/PRUZSKIVAJWSJDS7X7RAOQDTBB/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/PRUZSKIVAJWSJDS7X7RAOQDTBB/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":"2026-08-07T18:22:58Z","links":{"resolver":"https://pith.science/pith/PRUZSKIVAJWSJDS7X7RAOQDTBB","bundle":"https://pith.science/pith/PRUZSKIVAJWSJDS7X7RAOQDTBB/bundle.json","state":"https://pith.science/pith/PRUZSKIVAJWSJDS7X7RAOQDTBB/state.json","well_known_bundle":"https://pith.science/.well-known/pith/PRUZSKIVAJWSJDS7X7RAOQDTBB/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:PRUZSKIVAJWSJDS7X7RAOQDTBB","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":"97f220551d659f2e6279e4ccdbe7a3cdf4fa6c405ccbb2445683c8b75ce8b253","cross_cats_sorted":["cs.AI"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-07-04T05:46:20Z","title_canon_sha256":"a98d1f2e698caad6bacb002ff139954233293375d2ad5bbfa0c8508b8f29612d"},"schema_version":"1.0","source":{"id":"2407.03651","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2407.03651","created_at":"2026-07-05T08:43:29Z"},{"alias_kind":"arxiv_version","alias_value":"2407.03651v2","created_at":"2026-07-05T08:43:29Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2407.03651","created_at":"2026-07-05T08:43:29Z"},{"alias_kind":"pith_short_12","alias_value":"PRUZSKIVAJWS","created_at":"2026-07-05T08:43:29Z"},{"alias_kind":"pith_short_16","alias_value":"PRUZSKIVAJWSJDS7","created_at":"2026-07-05T08:43:29Z"},{"alias_kind":"pith_short_8","alias_value":"PRUZSKIV","created_at":"2026-07-05T08:43:29Z"}],"graph_snapshots":[{"event_id":"sha256:5000809b80d84a3f27d8cf0a75515dbdca0971d95496e151e5a7e1537a36ff86","target":"graph","created_at":"2026-07-05T08:43:29Z","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/2407.03651/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Large language models are prominently used in real-world applications, often tasked with reasoning over large volumes of documents. An exciting development in this space is models boasting extended context capabilities, with some accommodating over 2 million tokens. Such long context model capabilities remain uncertain in production systems, motivating the need to benchmark their performance on real world use cases. We address this challenge by proposing SWiM, an evaluation framework that addresses the limitations of standard tests. Testing the framework on eight long context models, we find t","authors_text":"Amanda Dsouza, Changho Shin, Christopher Glaze, Frederic Sala","cross_cats":["cs.AI"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-07-04T05:46:20Z","title":"Evaluating Language Model Context Windows: A \"Working Memory\" Test and Inference-time Correction"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2407.03651","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:a8306c8fb348daad45b59bda1e8b28178da34aa44a45e9213e0a6d9ea05840f1","target":"record","created_at":"2026-07-05T08:43:29Z","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":"97f220551d659f2e6279e4ccdbe7a3cdf4fa6c405ccbb2445683c8b75ce8b253","cross_cats_sorted":["cs.AI"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-07-04T05:46:20Z","title_canon_sha256":"a98d1f2e698caad6bacb002ff139954233293375d2ad5bbfa0c8508b8f29612d"},"schema_version":"1.0","source":{"id":"2407.03651","kind":"arxiv","version":2}},"canonical_sha256":"7c69992915026d248e5fbfe207407308427bd53678499f67a79c65b047ed4538","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"7c69992915026d248e5fbfe207407308427bd53678499f67a79c65b047ed4538","first_computed_at":"2026-07-05T08:43:29.908041Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T08:43:29.908041Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"2PZDoFVagY3fA1xevXj8/s1CV9rsMWSr+TPCQA8i/my75B8oh0g9ALUJROaDlf/AbaHm3HL//vL7w9JtoxdiAA==","signature_status":"signed_v1","signed_at":"2026-07-05T08:43:29.908511Z","signed_message":"canonical_sha256_bytes"},"source_id":"2407.03651","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:a8306c8fb348daad45b59bda1e8b28178da34aa44a45e9213e0a6d9ea05840f1","sha256:5000809b80d84a3f27d8cf0a75515dbdca0971d95496e151e5a7e1537a36ff86"],"state_sha256":"f254c174e60ea06699f41bd1249459f5856088f7ce03a5539d2801cfc485052e"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"cdPoSYjtf4MniZbK0NWmGcQ3Shfkt683IArJ1A8QCQmvWZvLR7qw5a29NhfO9CykQmSJFK23aBrwv8oXQ41rDw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-07T18:22:58.768096Z","bundle_sha256":"0a3d31258e571c34fcc43ddb89d742cea115da7cb281e99afd7be67e83235202"}}