{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:R5DA27DNW6KOV7C7QGR25WZGWX","short_pith_number":"pith:R5DA27DN","schema_version":"1.0","canonical_sha256":"8f460d7c6db794eafc5f81a3aedb26b5dc2883db34a0b76113854f7b519af3cf","source":{"kind":"arxiv","id":"2407.15504","version":2},"attestation_state":"computed","paper":{"title":"Fundamental Limits of Prompt Compression: A Rate-Distortion Framework for Black-Box Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL","cs.IT","math.IT"],"primary_cat":"cs.LG","authors_text":"Adway Girish, Alliot Nagle, Ashok Vardhan Makkuva, Hyeji Kim, Marco Bondaschi, Michael Gastpar","submitted_at":"2024-07-22T09:40:13Z","abstract_excerpt":"We formalize the problem of prompt compression for large language models (LLMs) and present a framework to unify token-level prompt compression methods which create hard prompts for black-box models. We derive the distortion-rate function for this setup as a linear program, and provide an efficient algorithm to compute this fundamental limit via the dual of the linear program. Using the distortion-rate function as the baseline, we study the performance of existing compression schemes on a synthetic dataset consisting of prompts generated from a Markov chain, natural language queries, and their"},"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":"2407.15504","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-07-22T09:40:13Z","cross_cats_sorted":["cs.CL","cs.IT","math.IT"],"title_canon_sha256":"1c8e61b3501c19610eb36ac6058a3130cc823e69f37d8d4c07fe6a0e316e51dd","abstract_canon_sha256":"9be2590a7a6df930aec3454a317f99c161466ef396b2ec4a26d950ef85db20b1"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:47:32.725887Z","signature_b64":"Brv/msSsmH6jtoqpOnFFYiQZgfjTras+cSmYKKYuFQTM6yN3oKP8jDQvMw33yuyGBsxFqxxlsZsY/ObXJV+IBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"8f460d7c6db794eafc5f81a3aedb26b5dc2883db34a0b76113854f7b519af3cf","last_reissued_at":"2026-07-05T09:47:32.725374Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:47:32.725374Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Fundamental Limits of Prompt Compression: A Rate-Distortion Framework for Black-Box Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL","cs.IT","math.IT"],"primary_cat":"cs.LG","authors_text":"Adway Girish, Alliot Nagle, Ashok Vardhan Makkuva, Hyeji Kim, Marco Bondaschi, Michael Gastpar","submitted_at":"2024-07-22T09:40:13Z","abstract_excerpt":"We formalize the problem of prompt compression for large language models (LLMs) and present a framework to unify token-level prompt compression methods which create hard prompts for black-box models. We derive the distortion-rate function for this setup as a linear program, and provide an efficient algorithm to compute this fundamental limit via the dual of the linear program. Using the distortion-rate function as the baseline, we study the performance of existing compression schemes on a synthetic dataset consisting of prompts generated from a Markov chain, natural language queries, and their"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2407.15504","kind":"arxiv","version":2},"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/2407.15504/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":"2407.15504","created_at":"2026-07-05T09:47:32.725438+00:00"},{"alias_kind":"arxiv_version","alias_value":"2407.15504v2","created_at":"2026-07-05T09:47:32.725438+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2407.15504","created_at":"2026-07-05T09:47:32.725438+00:00"},{"alias_kind":"pith_short_12","alias_value":"R5DA27DNW6KO","created_at":"2026-07-05T09:47:32.725438+00:00"},{"alias_kind":"pith_short_16","alias_value":"R5DA27DNW6KOV7C7","created_at":"2026-07-05T09:47:32.725438+00:00"},{"alias_kind":"pith_short_8","alias_value":"R5DA27DN","created_at":"2026-07-05T09:47:32.725438+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/R5DA27DNW6KOV7C7QGR25WZGWX","json":"https://pith.science/pith/R5DA27DNW6KOV7C7QGR25WZGWX.json","graph_json":"https://pith.science/api/pith-number/R5DA27DNW6KOV7C7QGR25WZGWX/graph.json","events_json":"https://pith.science/api/pith-number/R5DA27DNW6KOV7C7QGR25WZGWX/events.json","paper":"https://pith.science/paper/R5DA27DN"},"agent_actions":{"view_html":"https://pith.science/pith/R5DA27DNW6KOV7C7QGR25WZGWX","download_json":"https://pith.science/pith/R5DA27DNW6KOV7C7QGR25WZGWX.json","view_paper":"https://pith.science/paper/R5DA27DN","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2407.15504&json=true","fetch_graph":"https://pith.science/api/pith-number/R5DA27DNW6KOV7C7QGR25WZGWX/graph.json","fetch_events":"https://pith.science/api/pith-number/R5DA27DNW6KOV7C7QGR25WZGWX/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/R5DA27DNW6KOV7C7QGR25WZGWX/action/timestamp_anchor","attest_storage":"https://pith.science/pith/R5DA27DNW6KOV7C7QGR25WZGWX/action/storage_attestation","attest_author":"https://pith.science/pith/R5DA27DNW6KOV7C7QGR25WZGWX/action/author_attestation","sign_citation":"https://pith.science/pith/R5DA27DNW6KOV7C7QGR25WZGWX/action/citation_signature","submit_replication":"https://pith.science/pith/R5DA27DNW6KOV7C7QGR25WZGWX/action/replication_record"}},"created_at":"2026-07-05T09:47:32.725438+00:00","updated_at":"2026-07-05T09:47:32.725438+00:00"}