{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:GTX37APVYO6T3TUYQBXZNLAEPK","short_pith_number":"pith:GTX37APV","schema_version":"1.0","canonical_sha256":"34efbf81f5c3bd3dce98806f96ac047a8e3229e061afd52c6de713b24d41c129","source":{"kind":"arxiv","id":"2606.22953","version":1},"attestation_state":"computed","paper":{"title":"Plans Don't Persist: Why Context Management Is Load Bearing for LLM Agents","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.AI","authors_text":"Aman Mehta, Anupam Datta","submitted_at":"2026-06-22T07:30:47Z","abstract_excerpt":"Long-horizon agents depend on context management: systems compress, summarize, and evict old tokens so tasks can continue beyond finite windows. That is safe only when dropped information is no longer needed or has been internalized. Plans are the stress case: they are written early, used for many steps, and first to be evicted. We introduce replay pairing, a diagnostic that runs the same trajectory with and without the plan in history and measures hidden-state cosine distance. On Llama-3.1-70B, plan signal spikes to 0.453 one step after the plan, then falls 4.1x in a single action-observation"},"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":"2606.22953","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2026-06-22T07:30:47Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"8d8d718773714b49ebf50237ad390826c6dfe18e864afe6e19dd6197f15e769a","abstract_canon_sha256":"0d2c75be5eb63c267e4a208ea39f22ecda7b84183d3e0aa7c9984cb4bdc3b451"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-23T03:14:05.145006Z","signature_b64":"VhXl7GKLNRGDUFnxIzKGZXnq91++NQFRYngt30EOsRLt77IIj3RHCJZleESHPvd0AWDU50scD/tuPVCdBuTiBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"34efbf81f5c3bd3dce98806f96ac047a8e3229e061afd52c6de713b24d41c129","last_reissued_at":"2026-06-23T03:14:05.144654Z","signature_status":"signed_v1","first_computed_at":"2026-06-23T03:14:05.144654Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Plans Don't Persist: Why Context Management Is Load Bearing for LLM Agents","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.AI","authors_text":"Aman Mehta, Anupam Datta","submitted_at":"2026-06-22T07:30:47Z","abstract_excerpt":"Long-horizon agents depend on context management: systems compress, summarize, and evict old tokens so tasks can continue beyond finite windows. That is safe only when dropped information is no longer needed or has been internalized. Plans are the stress case: they are written early, used for many steps, and first to be evicted. We introduce replay pairing, a diagnostic that runs the same trajectory with and without the plan in history and measures hidden-state cosine distance. On Llama-3.1-70B, plan signal spikes to 0.453 one step after the plan, then falls 4.1x in a single action-observation"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.22953","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/2606.22953/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":"2606.22953","created_at":"2026-06-23T03:14:05.144710+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.22953v1","created_at":"2026-06-23T03:14:05.144710+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.22953","created_at":"2026-06-23T03:14:05.144710+00:00"},{"alias_kind":"pith_short_12","alias_value":"GTX37APVYO6T","created_at":"2026-06-23T03:14:05.144710+00:00"},{"alias_kind":"pith_short_16","alias_value":"GTX37APVYO6T3TUY","created_at":"2026-06-23T03:14:05.144710+00:00"},{"alias_kind":"pith_short_8","alias_value":"GTX37APV","created_at":"2026-06-23T03:14:05.144710+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/GTX37APVYO6T3TUYQBXZNLAEPK","json":"https://pith.science/pith/GTX37APVYO6T3TUYQBXZNLAEPK.json","graph_json":"https://pith.science/api/pith-number/GTX37APVYO6T3TUYQBXZNLAEPK/graph.json","events_json":"https://pith.science/api/pith-number/GTX37APVYO6T3TUYQBXZNLAEPK/events.json","paper":"https://pith.science/paper/GTX37APV"},"agent_actions":{"view_html":"https://pith.science/pith/GTX37APVYO6T3TUYQBXZNLAEPK","download_json":"https://pith.science/pith/GTX37APVYO6T3TUYQBXZNLAEPK.json","view_paper":"https://pith.science/paper/GTX37APV","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.22953&json=true","fetch_graph":"https://pith.science/api/pith-number/GTX37APVYO6T3TUYQBXZNLAEPK/graph.json","fetch_events":"https://pith.science/api/pith-number/GTX37APVYO6T3TUYQBXZNLAEPK/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/GTX37APVYO6T3TUYQBXZNLAEPK/action/timestamp_anchor","attest_storage":"https://pith.science/pith/GTX37APVYO6T3TUYQBXZNLAEPK/action/storage_attestation","attest_author":"https://pith.science/pith/GTX37APVYO6T3TUYQBXZNLAEPK/action/author_attestation","sign_citation":"https://pith.science/pith/GTX37APVYO6T3TUYQBXZNLAEPK/action/citation_signature","submit_replication":"https://pith.science/pith/GTX37APVYO6T3TUYQBXZNLAEPK/action/replication_record"}},"created_at":"2026-06-23T03:14:05.144710+00:00","updated_at":"2026-06-23T03:14:05.144710+00:00"}