{"paper":{"title":"Tuning for TraceTarnish: Techniques, Trends, and Testing Tangible Traits","license":"http://creativecommons.org/licenses/by/4.0/","headline":"TraceTarnish attack analysis identifies function-word frequencies, content-word distributions, and type-token ratio as reliable signals that text has been altered to mask its author.","cross_cats":["cs.CL","cs.IR"],"primary_cat":"cs.CR","authors_text":"Robert Dilworth","submitted_at":"2025-12-03T05:39:40Z","abstract_excerpt":"In this study, we more rigorously evaluated our attack script $\\textit{TraceTarnish}$, which leverages adversarial stylometry principles to anonymize the authorship of text-based messages. To ensure the efficacy and utility of our attack, we sourced, processed, and analyzed Reddit comments -- comments that were later alchemized into $\\textit{TraceTarnish}$ data -- to gain valuable insights. The transformed $\\textit{TraceTarnish}$ data was then further augmented by $\\textit{StyloMetrix}$ to manufacture stylometric features -- features that were culled using the Information Gain criterion, leavi"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"The identified stylometric cues — function-word frequencies, content-word distributions, and the Type-Token Ratio — serve as reliable indicators of compromise (IoCs), revealing when a text has been deliberately altered to mask its true author.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That the Information Gain-selected features remain useful for both attack enhancement and detection even when only the transformed text is available, and that the Reddit-derived dataset generalizes beyond the specific comments and transformations tested.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"TraceTarnish attack identifies stylometric features like function-word frequencies and type-token ratio that both strengthen authorship anonymization and serve as indicators of compromise when pre- and post-transformation texts can be compared.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"TraceTarnish attack analysis identifies function-word frequencies, content-word distributions, and type-token ratio as reliable signals that text has been altered to mask its author.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"96b8a4d0a28b9206485595b3f786dcfe2d52e3c15297bfe4ff6e0370f0ecc472"},"source":{"id":"2512.03465","kind":"arxiv","version":5},"verdict":{"id":"6318afe4-14f8-4627-95f7-ca5fb81ff400","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-17T02:54:59.777440Z","strongest_claim":"The identified stylometric cues — function-word frequencies, content-word distributions, and the Type-Token Ratio — serve as reliable indicators of compromise (IoCs), revealing when a text has been deliberately altered to mask its true author.","one_line_summary":"TraceTarnish attack identifies stylometric features like function-word frequencies and type-token ratio that both strengthen authorship anonymization and serve as indicators of compromise when pre- and post-transformation texts can be compared.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That the Information Gain-selected features remain useful for both attack enhancement and detection even when only the transformed text is available, and that the Reddit-derived dataset generalizes beyond the specific comments and transformations tested.","pith_extraction_headline":"TraceTarnish attack analysis identifies function-word frequencies, content-word distributions, and type-token ratio as reliable signals that text has been altered to mask its author."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2512.03465/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":1,"snapshot_sha256":"421dbda32a0fcb02d8155b446011a5b074e2b416f002331ed59b8b6acf949147"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"}