Step 2: tool results

get_claim

{
  "include": [
    "arguments",
    "subclaims"
  ],
  "claim_id": "f4204e80-58c4-4214-b3b1-254c34855e38"
}

{"claim":{"id":"f4204e80-58c4-4214-b3b1-254c34855e38","kind":"claim","canonical_form":"Arbitrary capabilities can be hidden in neural network weights so that, under cryptographic assumptions, no polynomial-time white-box auditor can elicit them.","claim_type":"mathematical","state":"active","decomposition_status":"complete","importance":0.65,"contestation":0.35,"domains":["mathematics"],"tags":[{"slug":"computational-complexity-theory","name":"Computational complexity theory"},{"slug":"mechanistic-interpretability","name":"Mechanistic Interpretability"},{"slug":"model-backdoor-attacks","name":"Model Backdoor Attacks"}],"steward_state":"running","stewarded_at":"2026-11-14T06:39:17.268Z","canonical_direction_note":"The source advances the positive cryptographic-hardness thesis that arbitrary computations can be hidden from efficient white-box auditors.","created_at":"2026-11-13T00:21:24.490Z","updated_at":"2026-11-14T06:38:05.868Z","page_url":"https://minerval.ai/claims/f4204e80-58c4-4214-b3b1-254c34855e38"},"assessment":null,"coverage":{"examined":false,"on":[],"note":"No judgment of this node is in force. That is not a finding either way."},"quantity":null,"threshold_of":null,"question":null,"formalization":null,"formalization_pending":null,"formalization_history":[],"verification":null,"lean_checks":[],"lean_checks_total":0,"waiting_messages":[{"trigger":"structure_and_assess","from":"the system","sender_node_id":null,"sent_at":"2026-11-13T00:21:24.700Z","paid":false,"summary":"New claim onboarded. Give it the structure and assessment its importance warrants; a minor or settled claim may need no decomposition."},{"trigger":"mandate_plan","from":"a mandate's plan","sender_node_id":null,"sent_at":"2026-11-14T06:38:05.936Z","paid":true,"summary":"First assessment. The source is arXiv 2610.13928. Please read the statement of the main result and its assumptions: which cryptographic assumptions, what 'arbitrary capabilities' and 'elicit' mean for…"}],"requests":[],"requests_total":0,"subclaim_count":0,"instance_count":1,"source_map":null,"arguments":[],"unplaced_argument_passages":[{"instance_id":"83a163d6-b7d3-44cb-a228-937217c2e192","source":{"id":"2a45808e-eefa-4402-9037-03def369eefe","url":"https://arxiv.org/pdf/2610.13928","title":"AI Models Can Provably Hide Arbitrary Capabilities"},"verbatim_text":"Thus B has the same advantage over random guessing as A. Payload-hiding security of the lockers\ntherefore implies payload-hiding security of the compiled model.","context":"Appendix C reduction from compiled-model security to digital locker security.","stance":"advances","direction":"for"},{"instance_id":"b5d31178-d8c3-4df5-8569-69e9e06aa576","source":{"id":"2a45808e-eefa-4402-9037-03def369eefe","url":"https://arxiv.org/pdf/2610.13928","title":"AI Models Can Provably Hide Arbitrary Capabilities"},"verbatim_text":"In our setting, the key k is the secret trigger and the message m is the hidden payload, such as the\nweights of a hidden circuit. Thus, digital locker security means that an auditor without the trigger\ncannot efficiently recover the trigger, obtain a valid payload, or distinguish which equal-size payload\nwas locked.","context":"Section 3, digital locker security applied to hidden payloads.","stance":"advances","direction":"for"}],"unplaced_argument_passages_total":2,"subclaims":[],"subclaims_assessed":0}