mlsanityData to 14 Oct 2026

Stochastic Semantic Evidence Graphs: Uncertainty Propagation and Governance for Agentic AI

arXiv 2609.29703 ↗First posted 3 Sep 2026cs.CL cs.IR stat.MLcited by 0

On the map: LLM agents and tool use → Agent threats, guardrails and governance

Abstract (arXiv)

AI-workflow governance cannot be reduced to checking the final answer: an apparently safe answer may rest on a fragile evidence path that ordinary evaluation cannot see, localize or govern. We call this hidden fragility a "structural iceberg": hallucinations and unsupported claims may form its visible tip, while consequential weakness remains submerged. Stochastic semantic evidence graphs (SSEGs) expose these icebergs by preserving workflow channels, propagating local uncertainty and identifying the hidden paths on which an apparently safe output depends. ALCE and RAGTruth show that visible failures at the tip -unsupported citations and hallucinated spans -rest on distinct submerged weaknesses and therefore require different interventions. Across retrieval, tool-use and controlled stress tests, SSEG localizes those weaknesses, supports targeted repair, produces no false automatic passes in 35,000 known-truth cases and reduces ToolSandbox review by 28.8% across 96 executions from two agent models. The same structural view carries into end-to-end governance: in a separately sealed 1,200-case FinGovBench study, adding SSEG to GPT-OSS-20B reduces unsafe releases from 452/660 to 8/660 while releasing all 540 safe cases and correctly distinguishing 592/600 matched workflow pairs. An unchanged-gate transfer to Qwen3-8B releases all 540 safe cases and none of 660 unsafe cases, whereas flat-UQ releases 520 unsafe cases. SSEG therefore moves governance below surface-level output checking, turning hidden evidence dependencies into auditable, path-specific decisions about intervention, revalidation and release.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in Agent threats, guardrails and governance, one of 16 directions in LLM agents and tool use. Builds on 34 of its 44 references that are in the map. First posted on arXiv 3 Sep 2026.

Builds on

34 of 44 references.

References · 34

  1. 2020Retrieval-Augmented Generation for Knowledge-Intensive NLP Taskscited by 9,297
  2. 2025Qwen3 Technical Reportcited by 8,820
  3. 2018HotpotQA: A Dataset for Diverse, Explainable Multi-hop Question Answeringcited by 3,754
  4. 2022Language Models (Mostly) Know What They Knowcited by 1,856
  5. 2023Self-RAG: Learning to Retrieve, Generate, and Critique through Self-Reflectioncited by 1,839
  6. 2019Can You Trust Your Model's Uncertainty? Evaluating Predictive Uncertainty Under Dataset Shiftcited by 1,287
  7. 2023FActScore: Fine-grained Atomic Evaluation of Factual Precision in Long Form Text Generationcited by 1,282
  8. 2018Model Cards for Model Reportingcited by 1,080
  9. 2021MuSiQue: Multihop Questions via Single-hop Question Compositioncited by 916
  10. 2023Semantic Uncertainty: Linguistic Invariances for Uncertainty Estimation in Natural Language Generationcited by 819
  11. 2023Enabling Large Language Models to Generate Text with Citationscited by 595
  12. 2020Fact or Fiction: Verifying Scientific Claimscited by 554
22 more
  1. 2023Ragas: Automated Evaluation of Retrieval Augmented Generationcited by 418
  2. 2020Closing the AI Accountability Gap: Defining an End-to-End Framework for Internal Algorithmic Auditingcited by 384
  3. 2019SelectiveNet: A Deep Neural Network with an Integrated Reject Optioncited by 350
  4. 2022Conformal Risk Controlcited by 299
  5. 2024ToolSandbox: A Stateful, Conversational, Interactive Evaluation Benchmark for LLM Tool Use Capabilitiescited by 295
  6. 2023RAGTruth: A Hallucination Corpus for Developing Trustworthy Retrieval-Augmented Language Modelscited by 283
  7. 2023ARES: An Automated Evaluation Framework for Retrieval-Augmented Generation Systemscited by 265
  8. 2021Get Your Vitamin C! Robust Fact Verification with Contrastive Evidencecited by 225
  9. 2025Deep Think with Confidencecited by 205
  10. 2024BALROG: Benchmarking Agentic LLM and VLM Reasoning On Gamescited by 146
  11. 2025SWE-rebench: An Automated Pipeline for Task Collection and Decontaminated Evaluation of Software Engineering Agentscited by 142
  12. 2025PROV-AGENT: Unified Provenance for Tracking AI Agent Interactions in Agentic Workflowscited by 31
  13. 2024RCAEval: A Benchmark for Root Cause Analysis of Microservice Systems with Telemetry Datacited by 29
  14. 2024Do LLMs estimate uncertainty well in instruction-following?cited by 26
  15. 2025DRAGged into Conflicts: Detecting and Addressing Conflicting Sources in Search-Augmented LLMscited by 25
  16. 2025Trained on Tokens, Calibrated on Concepts: The Emergence of Semantic Calibration in LLMscited by 19
  17. 2025MetaFaith: Faithful Natural Language Uncertainty Expression in LLMscited by 16
  18. 2025Attributing Response to Context: A Jensen-Shannon Divergence Driven Mechanistic Study of Context Attribution in Retrieval-Augmented Generationcited by 12
  19. 2025SGIC: A Self-Guided Iterative Calibration Framework for RAGcited by 8
  20. 2025MAGIC: A Multi-Hop and Graph-Based Benchmark for Inter-Context Conflicts in Retrieval-Augmented Generationcited by 6
  21. 2026Calibrating Semantic Uncertainty from Observable Language-Model Probabilitiescited by 4
  22. 2026Uncertainty Quantification for LLM Function-Callingcited by 2

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026GRUET: Quantifying Uncertainty of Agentic Reasoning-and-Acting Processes
  2. 2026Building Trustworthy Graph-Agentic RAG for Social Good: Architectures, Failure Propagation, and Assurance by Construction
  3. 2026MAP-Graph: Provenance-Aware Shared Memory for Multi-Agent Workflows
  4. 2024BayesAgent: Bayesian Agentic Reasoning Under Uncertainty via Verbalized Probabilistic Graphical Modeling
  5. 2026VeriGraph: Towards Verifiable Data-Analytic Agents
  6. 2026Semantic Bayesian World Models
  7. 2026EviGraph: Evidence-Guided Autonomous Research Agents
  8. 2026DNative-Twin: Decision Graphs and Digital Twins for Reconstructable Agentic Decisions