LLM Agent Reliability

LLM: Large Language Model

Overview

LLM agent reliability studies whether agents can complete tasks correctly, follow instructions and safety constraints, use evidence faithfully, and behave consistently across repeated or extended interactions. Common approaches combine structured harnesses and deterministic tool interfaces with state tracking, evidence checks, runtime monitoring, calibrated abstention, and recovery from failures; training methods also teach agents to verify actions or learn from execution feedback. Evaluations increasingly use interactive benchmarks, noisy or changing tools, repeated trials, and trajectory-level measures alongside final task success. A central tension is that added reasoning, delegation, memory, or orchestration can improve some tasks but also introduce new failure paths, so reliability depends on the model, harness, tools, and evaluation setting together.

Latest papers

CardsList

909 papers94 papers in the last four weeks, down 51% on the four weeks before. 0.7% of all new papers.

  1. DSFlow: A Neuro-Symbolic Data Workflow Agent with Statistical Diagnostic Guardrails

    Nov 18, 2026Junfeng Nie, Xiaohui ChenData Analysis AgentsLLM Agent Reliability

  2. Safe to Omit, Worth Omitting? Keeping the Answer and Cutting Latency When an LLM Agent Skips Another's Message

    Nov 17, 2026Xinyang Li, Chih-Hsuan Yang, Xiaolong Ma +4Multi-Agent LLM SystemsLLM Agent Efficiency

  3. Reliability by Architecture: Agentic Workflows for Validated Multi-Physics Simulation from Natural Language

    Nov 17, 2026Søren Taverniers, Eric Darve, Hamdi A. TchelepiAgentic WorkflowsLarge Language Model-Based Simulation

  4. Idempotency Is Not Recovery: A Fault-Injection Study of Language-Model Agent Tool Actions

    Nov 17, 2026Song XuTool-Using Language AgentsLLM Agent Reliability

  5. Can Coding Agents Write and Transpile Unit Tests

    Nov 16, 2026Gil Pasternak, AJ Langley, Tom Yotam +1AI Coding AgentsAutomated Test Generation

  6. System 1 agents: a method for local, reliable and fast agentic AI

    Nov 16, 2026Eleftherios Tzanis, Michail E. KlontzasLLM Agent ReliabilityLLM Tool Use

  7. Invariant-Aware Harness Enforcement to Protect LLM Agents from Silent API Failures

    Nov 16, 2026Sajib Acharjee Dip, Brian Ouyang, Lala MotlhabiRuntime Enforcement for AI AgentsLLM Agent Harnesses

  8. HydroMind: A Checkable Agent for Autonomous Hydrological Research

    Nov 15, 2026Zheng Fang, Yiqun Sun, Simin Qu +1LLM Agent HarnessesAutonomous Research Agents

  9. Does a completed simulation run mean a usable urban wind result for AI agents?

    Nov 14, 2026Zhongke Qu, Yunwei Zhang, Junwei Su +5LLM AgentsProcess-Based Agent Evaluation

  10. Beyond Clumsy Agentic Tool Calling: How Cueing and Spotting Make the Difference

    Nov 14, 2026Xindi Yang, Yan Liu, Jiajun SunTool-Using AgentsLLM Agents

  11. Auditable Authorization for LLM Agents: Empirical Evidence Across Model, Harness, and Workflow Layers

    Nov 14, 2026Bangjie Deng, Xiaoxuan Lu, Wenjun Chen +8AI Agent AuditingHuman Oversight of AI

  12. StateAlign: Internalizing State-Aware Decision-Making for Reliable Multi-Turn Tool Agents via Reinforcement Learning

    Nov 12, 2026Zhexi Lu, Subhajit Chaudhury, Tejaswini Pedapati +2LLM Agent ReliabilityRL for Tool Use

  13. Termination Lag: Coding Agents Solve Early but Stop Late

    Nov 11, 2026Musa Cim, Burak Topcu, Mahmut Taylan KandemirLLM AgentsLLM Agent Efficiency

  14. Cleaning an Agent's History Hurts When the World Still Holds the Error

    Nov 10, 2026Egor Pakhomov, Anurag Koul, Hongjun Liu +3LLM Agent Reliability

  15. Comparative Analysis of Agent Evaluation Frameworks: Stability, Detection, and Discovery in Enterprise Analytics Agents

    Nov 10, 2026Howard Zhang, Chuck Arvin, Logan Brown +3LLM Agent EvaluationLLM Agent Reliability

  16. When Should Agents Remember? Falsification-Gated Self-Evolution for LLM Agents

    Nov 10, 2026Runxuan Tang, Haoyu Gao, Yuyan Ding +4LLM Agent Self-ImprovementLLM Agent Memory

  17. Synapse-Control: Making Agent Experience Actionable with Lightweight Runtime Control

    Nov 10, 2026Zhiyuan Huang, Shengan Zheng, Chenghu Yang +5LLM AgentsLLM Agent Reliability

  18. Dependable Long-Horizon LLM Agents: A Trajectory-Centered Empirical Study of Where Errors Enter and Which Controls Reach Them

    Nov 9, 2026Shaohuang Wang, Jinghao Wen, Xulei Zhang +1Agent Failure AnalysisLLM Agent Reliability