{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:X7J73GIDH3IE2L5VHFI7A63RGQ","short_pith_number":"pith:X7J73GID","schema_version":"1.0","canonical_sha256":"bfd3fd99033ed04d2fb53951f07b7134103ab2cd9b5787f09e9c6aca58d53056","source":{"kind":"arxiv","id":"2506.17419","version":1},"attestation_state":"computed","paper":{"title":"UProp: Investigating the Uncertainty Propagation of LLMs in Multi-Step Agentic Decision-Making","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG","stat.ML"],"primary_cat":"cs.CL","authors_text":"Bhavya Kailkhura, James Diffenderfer, Jinhao Duan, Kaidi Xu, Sandeep Madireddy, Tianlong Chen","submitted_at":"2025-06-20T18:34:04Z","abstract_excerpt":"As Large Language Models (LLMs) are integrated into safety-critical applications involving sequential decision-making in the real world, it is essential to know when to trust LLM decisions. Existing LLM Uncertainty Quantification (UQ) methods are primarily designed for single-turn question-answering formats, resulting in multi-step decision-making scenarios, e.g., LLM agentic system, being underexplored. In this paper, we introduce a principled, information-theoretic framework that decomposes LLM sequential decision uncertainty into two parts: (i) internal uncertainty intrinsic to the current "},"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":"2506.17419","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-06-20T18:34:04Z","cross_cats_sorted":["cs.AI","cs.LG","stat.ML"],"title_canon_sha256":"c306a911112cab45cf536666350f3ec1cde60163e0367a2bc06a7e385c8238ed","abstract_canon_sha256":"ca37c8fc61d86dd81768b2f5f526eebc04c3462dc6a539f4175945227576d175"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:25:17.424260Z","signature_b64":"YhVufNs9Dua2OAvU2/tjGIfk+73HC2nWFrGf7M9z19W0s9GUXClp63gjNj8mnpaM7UNX1cWLCSqhYHu1fK3hAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"bfd3fd99033ed04d2fb53951f07b7134103ab2cd9b5787f09e9c6aca58d53056","last_reissued_at":"2026-07-05T11:25:17.423775Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:25:17.423775Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"UProp: Investigating the Uncertainty Propagation of LLMs in Multi-Step Agentic Decision-Making","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG","stat.ML"],"primary_cat":"cs.CL","authors_text":"Bhavya Kailkhura, James Diffenderfer, Jinhao Duan, Kaidi Xu, Sandeep Madireddy, Tianlong Chen","submitted_at":"2025-06-20T18:34:04Z","abstract_excerpt":"As Large Language Models (LLMs) are integrated into safety-critical applications involving sequential decision-making in the real world, it is essential to know when to trust LLM decisions. Existing LLM Uncertainty Quantification (UQ) methods are primarily designed for single-turn question-answering formats, resulting in multi-step decision-making scenarios, e.g., LLM agentic system, being underexplored. In this paper, we introduce a principled, information-theoretic framework that decomposes LLM sequential decision uncertainty into two parts: (i) internal uncertainty intrinsic to the current "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.17419","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/2506.17419/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":"2506.17419","created_at":"2026-07-05T11:25:17.423834+00:00"},{"alias_kind":"arxiv_version","alias_value":"2506.17419v1","created_at":"2026-07-05T11:25:17.423834+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.17419","created_at":"2026-07-05T11:25:17.423834+00:00"},{"alias_kind":"pith_short_12","alias_value":"X7J73GIDH3IE","created_at":"2026-07-05T11:25:17.423834+00:00"},{"alias_kind":"pith_short_16","alias_value":"X7J73GIDH3IE2L5V","created_at":"2026-07-05T11:25:17.423834+00:00"},{"alias_kind":"pith_short_8","alias_value":"X7J73GID","created_at":"2026-07-05T11:25:17.423834+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":9,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.19868","citing_title":"A Systematic Evaluation of Black-Box Uncertainty Estimation Methods for Large Language Models","ref_index":83,"is_internal_anchor":false},{"citing_arxiv_id":"2606.20662","citing_title":"Confidence Laundering in Agent Systems: Why Uncertainty Needs a Latent Carrier","ref_index":8,"is_internal_anchor":false},{"citing_arxiv_id":"2606.06976","citing_title":"Exploring Agentic Tool-Calling Decisions via Uncertainty-Aligned Reinforcement Learning","ref_index":1,"is_internal_anchor":false},{"citing_arxiv_id":"2607.00972","citing_title":"Bayesian Uncertainty Propagation for Agentic RAG Pipelines: A Proof-of-Concept Study on Multi-Hop Question Answering","ref_index":4,"is_internal_anchor":false},{"citing_arxiv_id":"2605.24756","citing_title":"Proper Scoring Rules for Agentic Uncertainty Quantification","ref_index":2,"is_internal_anchor":false},{"citing_arxiv_id":"2605.21801","citing_title":"Why Semantic Entropy Fails: Geometry-Aware and Calibrated Uncertainty for Policy Optimization","ref_index":19,"is_internal_anchor":false},{"citing_arxiv_id":"2605.20061","citing_title":"Rewarding Beliefs, Not Actions: Consistency-Guided Credit Assignment for Long-Horizon Agents","ref_index":10,"is_internal_anchor":false},{"citing_arxiv_id":"2604.23505","citing_title":"Uncertainty Propagation in LLM-Based Systems","ref_index":60,"is_internal_anchor":false},{"citing_arxiv_id":"2604.15109","citing_title":"IUQ: Interrogative Uncertainty Quantification for Long-Form Large Language Model Generation","ref_index":8,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/X7J73GIDH3IE2L5VHFI7A63RGQ","json":"https://pith.science/pith/X7J73GIDH3IE2L5VHFI7A63RGQ.json","graph_json":"https://pith.science/api/pith-number/X7J73GIDH3IE2L5VHFI7A63RGQ/graph.json","events_json":"https://pith.science/api/pith-number/X7J73GIDH3IE2L5VHFI7A63RGQ/events.json","paper":"https://pith.science/paper/X7J73GID"},"agent_actions":{"view_html":"https://pith.science/pith/X7J73GIDH3IE2L5VHFI7A63RGQ","download_json":"https://pith.science/pith/X7J73GIDH3IE2L5VHFI7A63RGQ.json","view_paper":"https://pith.science/paper/X7J73GID","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2506.17419&json=true","fetch_graph":"https://pith.science/api/pith-number/X7J73GIDH3IE2L5VHFI7A63RGQ/graph.json","fetch_events":"https://pith.science/api/pith-number/X7J73GIDH3IE2L5VHFI7A63RGQ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/X7J73GIDH3IE2L5VHFI7A63RGQ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/X7J73GIDH3IE2L5VHFI7A63RGQ/action/storage_attestation","attest_author":"https://pith.science/pith/X7J73GIDH3IE2L5VHFI7A63RGQ/action/author_attestation","sign_citation":"https://pith.science/pith/X7J73GIDH3IE2L5VHFI7A63RGQ/action/citation_signature","submit_replication":"https://pith.science/pith/X7J73GIDH3IE2L5VHFI7A63RGQ/action/replication_record"}},"created_at":"2026-07-05T11:25:17.423834+00:00","updated_at":"2026-07-05T11:25:17.423834+00:00"}