{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:32PNJDZVYXB7GTFY3SXTSHCVCG","short_pith_number":"pith:32PNJDZV","schema_version":"1.0","canonical_sha256":"de9ed48f35c5c3f34cb8dcaf391c55119a7863ad4516ac4d4d66aade7e3b3fe7","source":{"kind":"arxiv","id":"2607.04528","version":1},"attestation_state":"computed","paper":{"title":"Measuring Harness-Induced Belief Divergence in Multi-Step LLM Agents","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Haiwen Yi, Xinyuan Song","submitted_at":"2026-07-05T22:22:39Z","abstract_excerpt":"Software-agent benchmarks usually report whether an agent solves a task, but the agent reaches that outcome through a harness that controls what it sees, which actions it can take, which failures are repaired, which states are verified, and which evidence is logged. We show that this harness can change the agent's multi-step beliefs even when the task, environment, and base LLM are fixed. We introduce a belief-rollout diagnostic that elicits structured K-step trajectories over progress, risk, recoverability, constraints, failure mode, uncertainty, future success, repair cost, and next action u"},"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":"2607.04528","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2026-07-05T22:22:39Z","cross_cats_sorted":[],"title_canon_sha256":"12ca0e2bfc4cf7c2e75cd8ef6582ed5b3f20ee79b49a79ff0854e97eb4f488ea","abstract_canon_sha256":"85882d49c4c0de70f0e4130f79ade6851ad222429b27b7cf2106c8b788d9617c"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-07T02:19:23.321359Z","signature_b64":"67BZZjCK7ZHcrCYHEV+6tXiMAollbsKThy87Z6JfVUVCpAmuwaQusd/XTBIAbMmMpLacIp/jiZOeVakT8cUMBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"de9ed48f35c5c3f34cb8dcaf391c55119a7863ad4516ac4d4d66aade7e3b3fe7","last_reissued_at":"2026-07-07T02:19:23.320547Z","signature_status":"signed_v1","first_computed_at":"2026-07-07T02:19:23.320547Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Measuring Harness-Induced Belief Divergence in Multi-Step LLM Agents","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Haiwen Yi, Xinyuan Song","submitted_at":"2026-07-05T22:22:39Z","abstract_excerpt":"Software-agent benchmarks usually report whether an agent solves a task, but the agent reaches that outcome through a harness that controls what it sees, which actions it can take, which failures are repaired, which states are verified, and which evidence is logged. We show that this harness can change the agent's multi-step beliefs even when the task, environment, and base LLM are fixed. We introduce a belief-rollout diagnostic that elicits structured K-step trajectories over progress, risk, recoverability, constraints, failure mode, uncertainty, future success, repair cost, and next action u"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.04528","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/2607.04528/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":"2607.04528","created_at":"2026-07-07T02:19:23.320666+00:00"},{"alias_kind":"arxiv_version","alias_value":"2607.04528v1","created_at":"2026-07-07T02:19:23.320666+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.04528","created_at":"2026-07-07T02:19:23.320666+00:00"},{"alias_kind":"pith_short_12","alias_value":"32PNJDZVYXB7","created_at":"2026-07-07T02:19:23.320666+00:00"},{"alias_kind":"pith_short_16","alias_value":"32PNJDZVYXB7GTFY","created_at":"2026-07-07T02:19:23.320666+00:00"},{"alias_kind":"pith_short_8","alias_value":"32PNJDZV","created_at":"2026-07-07T02:19:23.320666+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/32PNJDZVYXB7GTFY3SXTSHCVCG","json":"https://pith.science/pith/32PNJDZVYXB7GTFY3SXTSHCVCG.json","graph_json":"https://pith.science/api/pith-number/32PNJDZVYXB7GTFY3SXTSHCVCG/graph.json","events_json":"https://pith.science/api/pith-number/32PNJDZVYXB7GTFY3SXTSHCVCG/events.json","paper":"https://pith.science/paper/32PNJDZV"},"agent_actions":{"view_html":"https://pith.science/pith/32PNJDZVYXB7GTFY3SXTSHCVCG","download_json":"https://pith.science/pith/32PNJDZVYXB7GTFY3SXTSHCVCG.json","view_paper":"https://pith.science/paper/32PNJDZV","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2607.04528&json=true","fetch_graph":"https://pith.science/api/pith-number/32PNJDZVYXB7GTFY3SXTSHCVCG/graph.json","fetch_events":"https://pith.science/api/pith-number/32PNJDZVYXB7GTFY3SXTSHCVCG/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/32PNJDZVYXB7GTFY3SXTSHCVCG/action/timestamp_anchor","attest_storage":"https://pith.science/pith/32PNJDZVYXB7GTFY3SXTSHCVCG/action/storage_attestation","attest_author":"https://pith.science/pith/32PNJDZVYXB7GTFY3SXTSHCVCG/action/author_attestation","sign_citation":"https://pith.science/pith/32PNJDZVYXB7GTFY3SXTSHCVCG/action/citation_signature","submit_replication":"https://pith.science/pith/32PNJDZVYXB7GTFY3SXTSHCVCG/action/replication_record"}},"created_at":"2026-07-07T02:19:23.320666+00:00","updated_at":"2026-07-07T02:19:23.320666+00:00"}