{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:THVPHY5FPNV6Z6JXCFHSAIKJGJ","short_pith_number":"pith:THVPHY5F","schema_version":"1.0","canonical_sha256":"99eaf3e3a57b6becf937114f202149326ca0b47b1c0f3fb094c060f104c93571","source":{"kind":"arxiv","id":"2607.09153","version":1},"attestation_state":"computed","paper":{"title":"KV-PRM: Efficient Process Reward Modeling via KV-Cache Transfer for Multi-Agent Test-Time Scaling","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Haibo Jin, Haohan Wang, Kaidi Xu, Peng Kuang, Xiaopeng Yuan, Xiaoyu Han, Yanli Wang, Ye Yu","submitted_at":"2026-07-10T07:16:43Z","abstract_excerpt":"Process Reward Models (PRMs) have been proven to be highly effective in guiding test-time scaling (TTS) methods, which significantly boost the capabilities of LLM-based multi-agent systems. However, existing PRMs are text-based: they re-encode the entire trajectory text from scratch. In long multi-agent rollouts, the scoring cost, growing quadratically with respect to sequence length L, creates a severe computational bottleneck, severely limiting PRMs' application in long-context scenarios. To resolve this, we introduce KV-PRM, a highly efficient process reward model that eliminates the heavy "},"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":"2607.09153","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2026-07-10T07:16:43Z","cross_cats_sorted":[],"title_canon_sha256":"cb8cff54fd97f370aa41c2d27c89fea089de84c6b12b813c8cd1f13f072d5548","abstract_canon_sha256":"c930164c00802c195f07d9fa44ee013f5a803b80f1f9bfce9ba9730d71143ee3"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-13T01:18:46.990411Z","signature_b64":"N0Q4vKtRHfMR/EfHR+oiZaeyceoAE8PKbgQxNm07NqlHkex6+AZX80Y26/Cf0tK53oIFo8uswRoL6waBL/qjCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"99eaf3e3a57b6becf937114f202149326ca0b47b1c0f3fb094c060f104c93571","last_reissued_at":"2026-07-13T01:18:46.989257Z","signature_status":"signed_v1","first_computed_at":"2026-07-13T01:18:46.989257Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"KV-PRM: Efficient Process Reward Modeling via KV-Cache Transfer for Multi-Agent Test-Time Scaling","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Haibo Jin, Haohan Wang, Kaidi Xu, Peng Kuang, Xiaopeng Yuan, Xiaoyu Han, Yanli Wang, Ye Yu","submitted_at":"2026-07-10T07:16:43Z","abstract_excerpt":"Process Reward Models (PRMs) have been proven to be highly effective in guiding test-time scaling (TTS) methods, which significantly boost the capabilities of LLM-based multi-agent systems. However, existing PRMs are text-based: they re-encode the entire trajectory text from scratch. In long multi-agent rollouts, the scoring cost, growing quadratically with respect to sequence length L, creates a severe computational bottleneck, severely limiting PRMs' application in long-context scenarios. To resolve this, we introduce KV-PRM, a highly efficient process reward model that eliminates the heavy "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.09153","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/2607.09153/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":"2607.09153","created_at":"2026-07-13T01:18:46.989868+00:00"},{"alias_kind":"arxiv_version","alias_value":"2607.09153v1","created_at":"2026-07-13T01:18:46.989868+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.09153","created_at":"2026-07-13T01:18:46.989868+00:00"},{"alias_kind":"pith_short_12","alias_value":"THVPHY5FPNV6","created_at":"2026-07-13T01:18:46.989868+00:00"},{"alias_kind":"pith_short_16","alias_value":"THVPHY5FPNV6Z6JX","created_at":"2026-07-13T01:18:46.989868+00:00"},{"alias_kind":"pith_short_8","alias_value":"THVPHY5F","created_at":"2026-07-13T01:18:46.989868+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/THVPHY5FPNV6Z6JXCFHSAIKJGJ","json":"https://pith.science/pith/THVPHY5FPNV6Z6JXCFHSAIKJGJ.json","graph_json":"https://pith.science/api/pith-number/THVPHY5FPNV6Z6JXCFHSAIKJGJ/graph.json","events_json":"https://pith.science/api/pith-number/THVPHY5FPNV6Z6JXCFHSAIKJGJ/events.json","paper":"https://pith.science/paper/THVPHY5F"},"agent_actions":{"view_html":"https://pith.science/pith/THVPHY5FPNV6Z6JXCFHSAIKJGJ","download_json":"https://pith.science/pith/THVPHY5FPNV6Z6JXCFHSAIKJGJ.json","view_paper":"https://pith.science/paper/THVPHY5F","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2607.09153&json=true","fetch_graph":"https://pith.science/api/pith-number/THVPHY5FPNV6Z6JXCFHSAIKJGJ/graph.json","fetch_events":"https://pith.science/api/pith-number/THVPHY5FPNV6Z6JXCFHSAIKJGJ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/THVPHY5FPNV6Z6JXCFHSAIKJGJ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/THVPHY5FPNV6Z6JXCFHSAIKJGJ/action/storage_attestation","attest_author":"https://pith.science/pith/THVPHY5FPNV6Z6JXCFHSAIKJGJ/action/author_attestation","sign_citation":"https://pith.science/pith/THVPHY5FPNV6Z6JXCFHSAIKJGJ/action/citation_signature","submit_replication":"https://pith.science/pith/THVPHY5FPNV6Z6JXCFHSAIKJGJ/action/replication_record"}},"created_at":"2026-07-13T01:18:46.989868+00:00","updated_at":"2026-07-13T01:18:46.989868+00:00"}