{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:RJT4PP26WJEETRJ5ASGPCLGQBX","short_pith_number":"pith:RJT4PP26","schema_version":"1.0","canonical_sha256":"8a67c7bf5eb24849c53d048cf12cd00dd66a4aa13451824097fb4b86fc189a29","source":{"kind":"arxiv","id":"2606.10304","version":1},"attestation_state":"computed","paper":{"title":"MIRAGE: A Polarity-Flipping Encoding Subspace in LLM Agents","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Chenguang Wang, Jihye Kim, Kargi Chauhan, Pratibha Revankar, Sadiba Nusrat Nur, Vincent Siu","submitted_at":"2026-06-09T01:45:19Z","abstract_excerpt":"When LLM agents are coerced into covertly encoding sensitive data (Base64, ROT13, acrostic, synonym chains, and beyond), the resulting outputs evade output-side detection but the underlying computation does not. Across nine encoding families and eight models from five architecture families, that computation is supported by a shared low-dimensional encoding subspace in the residual stream. A logistic-regression probe trained on eight encoding families recovers the held-out ninth at AUC 0.975-1.000, reading the computation rather than surface features. The same direction exhibits a second mechan"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2606.10304","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-06-09T01:45:19Z","cross_cats_sorted":[],"title_canon_sha256":"a7a4791c160a48f4fabffad7ce392378fa5e8f0da7c97015f4d343a4b5914506","abstract_canon_sha256":"2c6654f6b5087196c9b689dd11a26ace140c92fab4f2d9453a7d29a3fa82c0b2"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-10T01:10:10.749907Z","signature_b64":"uvMua+Apk/2WPf3iW368qtOBwAnix3jIBYGrfbjnRvWhl5mA2GEa9r8/CcPOJrr/6oE0zJJikMjs8At+HqlCAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"8a67c7bf5eb24849c53d048cf12cd00dd66a4aa13451824097fb4b86fc189a29","last_reissued_at":"2026-06-10T01:10:10.748919Z","signature_status":"signed_v1","first_computed_at":"2026-06-10T01:10:10.748919Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"MIRAGE: A Polarity-Flipping Encoding Subspace in LLM Agents","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Chenguang Wang, Jihye Kim, Kargi Chauhan, Pratibha Revankar, Sadiba Nusrat Nur, Vincent Siu","submitted_at":"2026-06-09T01:45:19Z","abstract_excerpt":"When LLM agents are coerced into covertly encoding sensitive data (Base64, ROT13, acrostic, synonym chains, and beyond), the resulting outputs evade output-side detection but the underlying computation does not. Across nine encoding families and eight models from five architecture families, that computation is supported by a shared low-dimensional encoding subspace in the residual stream. A logistic-regression probe trained on eight encoding families recovers the held-out ninth at AUC 0.975-1.000, reading the computation rather than surface features. The same direction exhibits a second mechan"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.10304","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/2606.10304/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"},"aliases":[{"alias_kind":"arxiv","alias_value":"2606.10304","created_at":"2026-06-10T01:10:10.749068+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.10304v1","created_at":"2026-06-10T01:10:10.749068+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.10304","created_at":"2026-06-10T01:10:10.749068+00:00"},{"alias_kind":"pith_short_12","alias_value":"RJT4PP26WJEE","created_at":"2026-06-10T01:10:10.749068+00:00"},{"alias_kind":"pith_short_16","alias_value":"RJT4PP26WJEETRJ5","created_at":"2026-06-10T01:10:10.749068+00:00"},{"alias_kind":"pith_short_8","alias_value":"RJT4PP26","created_at":"2026-06-10T01:10:10.749068+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/RJT4PP26WJEETRJ5ASGPCLGQBX","json":"https://pith.science/pith/RJT4PP26WJEETRJ5ASGPCLGQBX.json","graph_json":"https://pith.science/api/pith-number/RJT4PP26WJEETRJ5ASGPCLGQBX/graph.json","events_json":"https://pith.science/api/pith-number/RJT4PP26WJEETRJ5ASGPCLGQBX/events.json","paper":"https://pith.science/paper/RJT4PP26"},"agent_actions":{"view_html":"https://pith.science/pith/RJT4PP26WJEETRJ5ASGPCLGQBX","download_json":"https://pith.science/pith/RJT4PP26WJEETRJ5ASGPCLGQBX.json","view_paper":"https://pith.science/paper/RJT4PP26","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.10304&json=true","fetch_graph":"https://pith.science/api/pith-number/RJT4PP26WJEETRJ5ASGPCLGQBX/graph.json","fetch_events":"https://pith.science/api/pith-number/RJT4PP26WJEETRJ5ASGPCLGQBX/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/RJT4PP26WJEETRJ5ASGPCLGQBX/action/timestamp_anchor","attest_storage":"https://pith.science/pith/RJT4PP26WJEETRJ5ASGPCLGQBX/action/storage_attestation","attest_author":"https://pith.science/pith/RJT4PP26WJEETRJ5ASGPCLGQBX/action/author_attestation","sign_citation":"https://pith.science/pith/RJT4PP26WJEETRJ5ASGPCLGQBX/action/citation_signature","submit_replication":"https://pith.science/pith/RJT4PP26WJEETRJ5ASGPCLGQBX/action/replication_record"}},"created_at":"2026-06-10T01:10:10.749068+00:00","updated_at":"2026-06-10T01:10:10.749068+00:00"}