{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2023:QEIZBAJLGCXRLPB5K72QKOVDTO","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":"8a44738274fde9f0366ced7a895c076ed789e92271bcd4418c318e0c18d96edd","cross_cats_sorted":["cs.CL","cs.CR"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2023-12-05T14:12:15Z","title_canon_sha256":"67d6663a22e1d7c356c20f87d4972d221dddb6e189ffea61b0cfcbaee0ccb28d"},"schema_version":"1.0","source":{"id":"2312.02780","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2312.02780","created_at":"2026-07-05T07:20:42Z"},{"alias_kind":"arxiv_version","alias_value":"2312.02780v1","created_at":"2026-07-05T07:20:42Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2312.02780","created_at":"2026-07-05T07:20:42Z"},{"alias_kind":"pith_short_12","alias_value":"QEIZBAJLGCXR","created_at":"2026-07-05T07:20:42Z"},{"alias_kind":"pith_short_16","alias_value":"QEIZBAJLGCXRLPB5","created_at":"2026-07-05T07:20:42Z"},{"alias_kind":"pith_short_8","alias_value":"QEIZBAJL","created_at":"2026-07-05T07:20:42Z"}],"graph_snapshots":[{"event_id":"sha256:cfd1d6e28aa56eff6f7bbafd85eb11cefa95fcf22a5439cb19fb1b5dbf8850b3","target":"graph","created_at":"2026-07-05T07:20:42Z","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/2312.02780/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"We explore a class of adversarial attacks targeting the activations of language models. By manipulating a relatively small subset of model activations, $a$, we demonstrate the ability to control the exact prediction of a significant number (in some cases up to 1000) of subsequent tokens $t$. We empirically verify a scaling law where the maximum number of target tokens $t_\\mathrm{max}$ predicted depends linearly on the number of tokens $a$ whose activations the attacker controls as $t_\\mathrm{max} = \\kappa a$. We find that the number of bits of control in the input space needed to control a sin","authors_text":"Stanislav Fort","cross_cats":["cs.CL","cs.CR"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2023-12-05T14:12:15Z","title":"Scaling Laws for Adversarial Attacks on Language Model Activations"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2312.02780","kind":"arxiv","version":1},"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":"sha256:2987a51bdad4873f7b38820fae47a2e01ee818985d8f0b32d9e498de6d7f4edd","target":"record","created_at":"2026-07-05T07:20:42Z","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":"8a44738274fde9f0366ced7a895c076ed789e92271bcd4418c318e0c18d96edd","cross_cats_sorted":["cs.CL","cs.CR"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2023-12-05T14:12:15Z","title_canon_sha256":"67d6663a22e1d7c356c20f87d4972d221dddb6e189ffea61b0cfcbaee0ccb28d"},"schema_version":"1.0","source":{"id":"2312.02780","kind":"arxiv","version":1}},"canonical_sha256":"811190812b30af15bc3d57f5053aa39b8c70de9074a905f552d9b8c6b03eb237","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"811190812b30af15bc3d57f5053aa39b8c70de9074a905f552d9b8c6b03eb237","first_computed_at":"2026-07-05T07:20:42.928589Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T07:20:42.928589Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"MriPQU9Z3jxARx/73WILNfLJYToYO9ehJYCUvEl5cbIUzAVOmno7IdnJL2jNp1909nAt3JNuhnDYdVbT1z0QDQ==","signature_status":"signed_v1","signed_at":"2026-07-05T07:20:42.928998Z","signed_message":"canonical_sha256_bytes"},"source_id":"2312.02780","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:2987a51bdad4873f7b38820fae47a2e01ee818985d8f0b32d9e498de6d7f4edd","sha256:cfd1d6e28aa56eff6f7bbafd85eb11cefa95fcf22a5439cb19fb1b5dbf8850b3"],"state_sha256":"e21637bb5d47bdf084375a48c25e6eb7817663b88d702029f27fd7caf72053fc"}