{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:XY2G3K54XRXGW4UQENJRW7XC2Y","merge_version":"pith-open-graph-merge-v1","event_count":3,"valid_event_count":3,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"b34a9ab559bfbe0809b5076fad39b88a247e809cca6c95fde4dbc9bb224f482b","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.SE","submitted_at":"2026-05-17T12:24:11Z","title_canon_sha256":"5b4970652583941957cfafa8246885efc42f035cc19195e08dc4407050e0c755"},"schema_version":"1.0","source":{"id":"2605.17416","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.17416","created_at":"2026-05-20T00:03:57Z"},{"alias_kind":"arxiv_version","alias_value":"2605.17416v1","created_at":"2026-05-20T00:03:57Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.17416","created_at":"2026-05-20T00:03:57Z"},{"alias_kind":"pith_short_12","alias_value":"XY2G3K54XRXG","created_at":"2026-05-20T00:03:57Z"},{"alias_kind":"pith_short_16","alias_value":"XY2G3K54XRXGW4UQ","created_at":"2026-05-20T00:03:57Z"},{"alias_kind":"pith_short_8","alias_value":"XY2G3K54","created_at":"2026-05-20T00:03:57Z"}],"graph_snapshots":[{"event_id":"sha256:cdda720365cee2706b7bb5d9552db5e80cfaaa6686981a9abec3af021001edda","target":"graph","created_at":"2026-05-20T00:03:57Z","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":"Under this favorable target-file scaffold, systems-specific prompting yields only six target matches across 54 counted attempts."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"The manual target-matching rubric correctly identifies rediscovery of the intended Mythos bug without bias from the specific language or details omitted from the prompts."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"A benchmarking experiment finds low rediscovery rates for three models on six Mythos-linked bug tasks, with only six target matches across 54 attempts under controlled prompting."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"Even when supplied with the exact target source files, large language models rediscover only six of the intended Mythos-linked systems bugs across 54 attempts."}],"snapshot_sha256":"8bc23607f6ae3c44af980564a5d0efed0dd77c0bf6791d402556fbd233608c99"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":false,"detectors_run":[{"findings_count":0,"name":"doi_title_agreement","ran_at":"2026-05-19T23:31:19.978554Z","status":"completed","version":"1.0.0"},{"findings_count":1,"name":"doi_compliance","ran_at":"2026-05-19T23:21:01.317386Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"claim_evidence","ran_at":"2026-05-19T21:41:57.741798Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"ai_meta_artifact","ran_at":"2026-05-19T21:33:23.687165Z","status":"skipped","version":"1.0.0"}],"endpoint":"/pith/2605.17416/integrity.json","findings":[{"audited_at":"2026-05-19T23:21:01.317386Z","detected_arxiv_id":null,"detected_doi":"10.1109/TDSC.2021.3051525.11","detector":"doi_compliance","finding_type":"recoverable_identifier","note":"DOI in the printed bibliography is fragmented by whitespace or line breaks. A longer candidate (10.1109/TDSC.2021.3051525.11) was visible in the surrounding text but could not be confirmed against doi.org as printed.","ref_index":30,"severity":"advisory","verdict_class":"incontrovertible"}],"snapshot_sha256":"61fba97be5b9b0e639fa338d9da22bc687200000f9dcfa68c8475ae2863bf183","summary":{"advisory":1,"by_detector":{"doi_compliance":{"advisory":1,"critical":0,"informational":0,"total":1}},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Anthropic's April 2026 Mythos materials combine benchmark claims with concrete bug-finding stories across OpenBSD, FreeBSD, Linux, FFmpeg, and browsers. This paper reports a controlled target-file rediscovery experiment on six public or high-confidence Mythos-linked systems tasks. Each model receives the same target file or files, read-only source tools, three repeats per task, and one manual target-matching rubric; prompts omit CVE identifiers, patch hashes, advisory text, author names, disclosure dates, and answer key root cause language. The experiment contains 54 counted model-task attempt","authors_text":"Arthur Gervais, Isaac David","cross_cats":["cs.AI"],"headline":"Even when supplied with the exact target source files, large language models rediscover only six of the intended Mythos-linked systems bugs across 54 attempts.","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.SE","submitted_at":"2026-05-17T12:24:11Z","title":"Benchmarking Mythos-Linked Bug Rediscovery"},"references":{"count":55,"internal_anchors":2,"resolved_work":55,"sample":[{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":1,"title":"Partnering with mozilla to improve firefox’s security, 2026","work_id":"701e45ac-d338-45f4-a344-c82060f08323","year":2026},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":2,"title":"Anthropic. Project glasswing, 2026. URL https://www.anthropic.com/glasswing. Ac- cessed 2026-04-18","work_id":"1f8988a9-4dcd-4c22-a8ca-95feea19fc83","year":2026},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":3,"title":"Claude mythos preview system card, 2026","work_id":"369aa627-6d39-42c5-b6ce-2db6bcccffb5","year":2026},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":4,"title":"Assessing claude mythos preview’s cybersecurity capabilities,","work_id":"c886aa75-9505-4523-b5a5-3312a0281eef","year":null},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":5,"title":"Published 2026-04- 07; accessed 2026-04-18","work_id":"ef53a326-086f-4392-ba25-64a01a748b7d","year":2026}],"snapshot_sha256":"35f8ea7a036d81b0df234a7200d182f9fe12707271125e20f81fc5c7df45b720"},"source":{"id":"2605.17416","kind":"arxiv","version":1},"verdict":{"created_at":"2026-05-19T23:11:04.576523Z","id":"a08a541d-12ed-4c47-82f6-d3d932f7fa11","model_set":{"reader":"grok-4.3"},"one_line_summary":"A benchmarking experiment finds low rediscovery rates for three models on six Mythos-linked bug tasks, with only six target matches across 54 attempts under controlled prompting.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"Even when supplied with the exact target source files, large language models rediscover only six of the intended Mythos-linked systems bugs across 54 attempts.","strongest_claim":"Under this favorable target-file scaffold, systems-specific prompting yields only six target matches across 54 counted attempts.","weakest_assumption":"The manual target-matching rubric correctly identifies rediscovery of the intended Mythos bug without bias from the specific language or details omitted from the prompts."}},"verdict_id":"a08a541d-12ed-4c47-82f6-d3d932f7fa11"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:acd9285f97792c71f3e7aedb12682cac444b594aadadba1b323274d3f2f6a3b4","target":"record","created_at":"2026-05-20T00:03:57Z","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":"b34a9ab559bfbe0809b5076fad39b88a247e809cca6c95fde4dbc9bb224f482b","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.SE","submitted_at":"2026-05-17T12:24:11Z","title_canon_sha256":"5b4970652583941957cfafa8246885efc42f035cc19195e08dc4407050e0c755"},"schema_version":"1.0","source":{"id":"2605.17416","kind":"arxiv","version":1}},"canonical_sha256":"be346dabbcbc6e6b729023531b7ee2d60b7a933fd4b65260469b948487112420","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"be346dabbcbc6e6b729023531b7ee2d60b7a933fd4b65260469b948487112420","first_computed_at":"2026-05-20T00:03:57.341680Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-20T00:03:57.341680Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"PeDCscWjNgTi80oijjSg6W54ljTU7scPqOL9N5RisRS1yOjwEmS9Bg0vFYGNFi7MmfOL2fbAoIS0fo6GKF35Cw==","signature_status":"signed_v1","signed_at":"2026-05-20T00:03:57.343387Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.17416","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:51ac499be239681d170ad2b063fcf7f7467de9baff26f0f019d2306c46e5ae2b","sha256:acd9285f97792c71f3e7aedb12682cac444b594aadadba1b323274d3f2f6a3b4","sha256:cdda720365cee2706b7bb5d9552db5e80cfaaa6686981a9abec3af021001edda"],"state_sha256":"9853f5048e8cd7a2d0f92b1860ac535028df3c9fc2da0807d49762ebc14eb9a1"}