{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:JMFSU3WM2GYZHQWLWYZRUX4RCK","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":"c4b4745ff40a3882c510ccaa98166f9ae2db476d60f30047983682b4370c73a7","cross_cats_sorted":["math.IT"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.IT","submitted_at":"2026-05-15T23:46:22Z","title_canon_sha256":"9fa072348073860d6117401e6c15035fb9087e3f17baeb48fbdf6f68db980f80"},"schema_version":"1.0","source":{"id":"2605.16709","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.16709","created_at":"2026-05-20T00:02:37Z"},{"alias_kind":"arxiv_version","alias_value":"2605.16709v1","created_at":"2026-05-20T00:02:37Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.16709","created_at":"2026-05-20T00:02:37Z"},{"alias_kind":"pith_short_12","alias_value":"JMFSU3WM2GYZ","created_at":"2026-05-20T00:02:37Z"},{"alias_kind":"pith_short_16","alias_value":"JMFSU3WM2GYZHQWL","created_at":"2026-05-20T00:02:37Z"},{"alias_kind":"pith_short_8","alias_value":"JMFSU3WM","created_at":"2026-05-20T00:02:37Z"}],"graph_snapshots":[{"event_id":"sha256:f1e87994d94053d6da8687ad870a0414a3089a61e4d625084d7fc558428bdb13","target":"graph","created_at":"2026-05-20T00:02:37Z","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":"We study the information-theoretic limits of the model by combining Gelfand-Pinsker and channel synthesis coding techniques and obtain an exact characterization of the capacity. Our algorithm achieves a bit-error rate below 10 percent with a rate of 0.375 bits/token over short token lengths with negligible perplexity and distortion degradation."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"The formulation assumes the encoder has limited non-causal access to token distributions within each block and that LLM cover statistics are known and can be leveraged for covert embedding; if this access or knowledge is unavailable in fully causal real-world sampling, the capacity characterization and performance claims would not apply."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"Characterizes the exact capacity of multi-bit covert LLM watermarking via Gelfand-Pinsker and channel synthesis, then gives a polar-code algorithm achieving 0.375 bits/token at under 10% BER with negligible perplexity impact."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"Multi-bit covert watermarking for LLMs has an exactly characterized capacity that supports practical embedding at 0.375 bits per token."}],"snapshot_sha256":"190b28217324e749ac6618e98c9efbb20c774b968b0c1480fc386e63ad3e633b"},"formal_canon":{"evidence_count":2,"snapshot_sha256":"4e57282d7070d4b170a6277eebf6ee4fdb62e9f9d2393fccbbb3e2aa521adc39"},"integrity":{"available":true,"clean":true,"detectors_run":[{"findings_count":0,"name":"doi_compliance","ran_at":"2026-05-19T20:32:32.044280Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"doi_title_agreement","ran_at":"2026-05-19T20:31:19.210416Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"claim_evidence","ran_at":"2026-05-19T19:01:56.360689Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"ai_meta_artifact","ran_at":"2026-05-19T18:33:26.483087Z","status":"skipped","version":"1.0.0"}],"endpoint":"/pith/2605.16709/integrity.json","findings":[],"snapshot_sha256":"b63fc0d71e1c74275c7c8ff6d1c98aee3c8b37d82edbe0d5e6de56bf42a07a39","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"We study the problem of multi-bit watermarking for large language models (LLMs). We introduce a block-autoregressive model inspired by multi-token prediction, in which the encoder has limited non-causal access to token distributions within each block. This formulation enables an information-theoretic characterization of multi-bit watermarking capacity, by which the knowledge of LLM cover statistics is leveraged to enable a multi-bit covert embedding. We study the information-theoretic limits of the model by combining Gelfand-Pinsker and channel synthesis coding techniques and obtain an exact c","authors_text":"Matthieu R. Bloch, Sidong Guo, Teodora Baluta, Tyler Kann","cross_cats":["math.IT"],"headline":"Multi-bit covert watermarking for LLMs has an exactly characterized capacity that supports practical embedding at 0.375 bits per token.","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.IT","submitted_at":"2026-05-15T23:46:22Z","title":"Covert Multi-bit LLM Watermarking: An Information Theory and Coding Approach"},"references":{"count":46,"internal_anchors":3,"resolved_work":46,"sample":[{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":1,"title":"A watermark for large language models,","work_id":"daa02d2d-650a-4faa-b61e-fc4dc5ace0a6","year":2023},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":2,"title":"Undetectable watermarks for language models,","work_id":"21ab8221-481a-447a-9e38-bd0702708b96","year":2024},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":3,"title":"Heavywater and simplexwater: Distortion- free llm watermarks for low-entropy distributions,","work_id":"38348e92-4c65-4c63-b36c-4513a67f56cd","year":2025},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":4,"title":"Provable robust watermarking for AI-generated text","work_id":"ad6da7f1-2fd9-4768-bbd8-0568dc11215f","year":2023},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":5,"title":"Robust content-dependent high- fidelity watermark for tracking in digital cinema,","work_id":"b131c990-560b-4a55-9035-420a5748c73d","year":2003}],"snapshot_sha256":"d89bb85282acd8290cd4ab471478f2a27a5e9a1eb960c0dee22f3cac794cca88"},"source":{"id":"2605.16709","kind":"arxiv","version":1},"verdict":{"created_at":"2026-05-19T20:19:36.931891Z","id":"06c5a49d-0dc0-4a0d-b094-8a88ed0f1b74","model_set":{"reader":"grok-4.3"},"one_line_summary":"Characterizes the exact capacity of multi-bit covert LLM watermarking via Gelfand-Pinsker and channel synthesis, then gives a polar-code algorithm achieving 0.375 bits/token at under 10% BER with negligible perplexity impact.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"Multi-bit covert watermarking for LLMs has an exactly characterized capacity that supports practical embedding at 0.375 bits per token.","strongest_claim":"We study the information-theoretic limits of the model by combining Gelfand-Pinsker and channel synthesis coding techniques and obtain an exact characterization of the capacity. Our algorithm achieves a bit-error rate below 10 percent with a rate of 0.375 bits/token over short token lengths with negligible perplexity and distortion degradation.","weakest_assumption":"The formulation assumes the encoder has limited non-causal access to token distributions within each block and that LLM cover statistics are known and can be leveraged for covert embedding; if this access or knowledge is unavailable in fully causal real-world sampling, the capacity characterization and performance claims would not apply."}},"verdict_id":"06c5a49d-0dc0-4a0d-b094-8a88ed0f1b74"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:6059a39263933950acb9076ab1c8f39ecfbaf2032570501562ff3d7d92c511b7","target":"record","created_at":"2026-05-20T00:02:37Z","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":"c4b4745ff40a3882c510ccaa98166f9ae2db476d60f30047983682b4370c73a7","cross_cats_sorted":["math.IT"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.IT","submitted_at":"2026-05-15T23:46:22Z","title_canon_sha256":"9fa072348073860d6117401e6c15035fb9087e3f17baeb48fbdf6f68db980f80"},"schema_version":"1.0","source":{"id":"2605.16709","kind":"arxiv","version":1}},"canonical_sha256":"4b0b2a6eccd1b193c2cbb6331a5f91129660c483a188377524ed8b7201fd24fc","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"4b0b2a6eccd1b193c2cbb6331a5f91129660c483a188377524ed8b7201fd24fc","first_computed_at":"2026-05-20T00:02:37.673341Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-20T00:02:37.673341Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"gNQVyDOJHYdiiAxQngKm4ViyJ6H8uft99RIZzcCsuDizNOTXXIDMhhGx18k+V1IygEY2UirKEStDz4w3plpRCg==","signature_status":"signed_v1","signed_at":"2026-05-20T00:02:37.674217Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.16709","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:6059a39263933950acb9076ab1c8f39ecfbaf2032570501562ff3d7d92c511b7","sha256:f1e87994d94053d6da8687ad870a0414a3089a61e4d625084d7fc558428bdb13"],"state_sha256":"612c53b562973c328e6c0747ee13282bbf7a618e823a1e3818e00dc6e929568d"}