{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:IGPFAYCXFTYJWAYOSLHXOQ4TAP","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":"3122410118920069546efceb4c409f16fc4600b24e7703a8fdcb3fc53e8ff3f0","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2026-02-18T07:49:47Z","title_canon_sha256":"635d77fbd1eda53c4cd6f4a49799b35c1f57cb644153e4e92193aa1e23a02da9"},"schema_version":"1.0","source":{"id":"2602.16246","kind":"arxiv","version":3}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2602.16246","created_at":"2026-05-18T02:44:31Z"},{"alias_kind":"arxiv_version","alias_value":"2602.16246v3","created_at":"2026-05-18T02:44:31Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2602.16246","created_at":"2026-05-18T02:44:31Z"},{"alias_kind":"pith_short_12","alias_value":"IGPFAYCXFTYJ","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_16","alias_value":"IGPFAYCXFTYJWAYO","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_8","alias_value":"IGPFAYCX","created_at":"2026-05-18T12:33:37Z"}],"graph_snapshots":[{"event_id":"sha256:3064301ac411e077bf22a560695d3b5ae8489f9ecdfad4d786b44429e9485c73","target":"graph","created_at":"2026-05-18T02:44:31Z","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":"Proxy state-based evaluation offers a practical, scalable alternative to deterministic agentic benchmarks for industrial LLM agents, producing stable model-differentiating rankings and on-/off-policy supervision that transfers to unseen scenarios."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"That LLM state trackers and judges, when given carefully specified scenarios, can infer accurate proxy states and verify goal completion with near-zero hallucination rates and high reliability."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"Introduces Proxy State-Based Evaluation as a scalable LLM-based method for verifiable assessment of multi-turn tool-calling agents using proxy states inferred from interaction traces."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"Proxy state-based evaluation replaces costly deterministic backends with LLM trackers and judges for benchmarking multi-turn tool-calling agents."}],"snapshot_sha256":"286af90951987f52f1120a23f176b998232573cdf0375274e7c2e03f731d5e4d"},"formal_canon":{"evidence_count":2,"snapshot_sha256":"0797ce130752c9cd8aeb36108b5c782c9e8a57972f90fd47d2f697c1296e5f75"},"paper":{"abstract_excerpt":"Interactive large language model (LLM) agents operating via multi-turn dialogue and multi-step tool calling are increasingly used in production. Benchmarks for these agents must both reliably compare models and yield on-policy training data. Prior agentic benchmarks, such as tau-bench, tau^2-bench, and AppWorld, rely on fully deterministic backends, which are costly to build and iterate. We propose Proxy State-Based Evaluation, an LLM-driven simulation framework that preserves final state-based evaluation without a deterministic database. Specifically, a scenario specifies the user goal, user/","authors_text":"Alec Chiu, Avinash Thangali, Chaitanya Kulkarni, Linsey Pang, Prakhar Mehrotra, Shivani Shekhar, Uma Kona, Yirou Ge, Yixi Li, Yun-Shiuan Chuang, Zijie Pan","cross_cats":[],"headline":"Proxy state-based evaluation replaces costly deterministic backends with LLM trackers and judges for benchmarking multi-turn tool-calling agents.","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2026-02-18T07:49:47Z","title":"Toward Scalable Verifiable Reward: Proxy State-Based Evaluation for Multi-turn Tool-Calling LLM Agents"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2602.16246","kind":"arxiv","version":3},"verdict":{"created_at":"2026-05-15T21:40:51.673070Z","id":"8a08f2dd-1f4c-4ad6-ae6b-6b020ed52f0f","model_set":{"reader":"grok-4.3"},"one_line_summary":"Introduces Proxy State-Based Evaluation as a scalable LLM-based method for verifiable assessment of multi-turn tool-calling agents using proxy states inferred from interaction traces.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"Proxy state-based evaluation replaces costly deterministic backends with LLM trackers and judges for benchmarking multi-turn tool-calling agents.","strongest_claim":"Proxy state-based evaluation offers a practical, scalable alternative to deterministic agentic benchmarks for industrial LLM agents, producing stable model-differentiating rankings and on-/off-policy supervision that transfers to unseen scenarios.","weakest_assumption":"That LLM state trackers and judges, when given carefully specified scenarios, can infer accurate proxy states and verify goal completion with near-zero hallucination rates and high reliability."}},"verdict_id":"8a08f2dd-1f4c-4ad6-ae6b-6b020ed52f0f"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:c6c9cf56bc6294368ef83da982ac2e83d8d3490316f2899db55033d1b2a2b6d7","target":"record","created_at":"2026-05-18T02:44:31Z","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":"3122410118920069546efceb4c409f16fc4600b24e7703a8fdcb3fc53e8ff3f0","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2026-02-18T07:49:47Z","title_canon_sha256":"635d77fbd1eda53c4cd6f4a49799b35c1f57cb644153e4e92193aa1e23a02da9"},"schema_version":"1.0","source":{"id":"2602.16246","kind":"arxiv","version":3}},"canonical_sha256":"419e5060572cf09b030e92cf77439303ed3a8dde833cd41f12183ab4a2884b89","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"419e5060572cf09b030e92cf77439303ed3a8dde833cd41f12183ab4a2884b89","first_computed_at":"2026-05-18T02:44:31.143396Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-18T02:44:31.143396Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"HFJQtDPPS7c+Sm+S1YYUDRORrVefBlxLBMirhPIIoWI2LcchUtRsa5TD761RA4XUdmfdr2VxxDV/iym1FoBXBA==","signature_status":"signed_v1","signed_at":"2026-05-18T02:44:31.143835Z","signed_message":"canonical_sha256_bytes"},"source_id":"2602.16246","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:c6c9cf56bc6294368ef83da982ac2e83d8d3490316f2899db55033d1b2a2b6d7","sha256:3064301ac411e077bf22a560695d3b5ae8489f9ecdfad4d786b44429e9485c73"],"state_sha256":"9cb3098d54bd565334ac497aa6888aaa5bbbac2d9e2347e0a2a973aebf2c5b60"}