AgentAbstain: Do LLM Agents Know When Not to Act?
2026-07-11 • Artificial Intelligence
Artificial Intelligence
AI summaryⓘ
The authors studied how well AI agents based on large language models can decide when to take action and when to abstain or not act. They created a new test called AgentAbstain, which has paired tasks showing when agents should act versus when they should not, across different tricky situations. They also built an automatic system, AbstainGen, to make many such tasks reliably. Testing 17 different AI agents showed even the best agent struggled to correctly handle both acting and abstaining, and being good at tasks didn’t mean being good at knowing when not to act. The authors highlight risks of agents acting wrongly before realizing they should hold back and provide their tools for others to use.
large language modelsagent abstentiontask evaluationsandbox environmentspaired-task benchmarktool-using agentstask successautonomous agentsautomatic task generationfailure modes
Authors
Xun Liu, Yi Evie Zhang, Vira Kasprova, Parisa Rabbani, Pardis Sadat Zahraei, Tianyu Zhang, Ali Ebrahimpour-Boroojeny, Varun Chandrasekaran
Abstract
Agent systems based on large language models (LLMs) are increasingly deployed for autonomous tasks, yet existing evaluations mostly focus on task success rather than whether agents know when to abstain. This gap poses real risks: under ambiguity, conflicting constraints, or tool failures, agents may execute unintended and irreversible actions. To close this gap, we present the first systematic evaluation framework for agentic abstention: the calibrated ability of tool-using LLM agents to recognize when not to act. At its core, AgentAbstain is a paired-task benchmark built on an agent-native taxonomy of 8 abstention scenarios across pre-execution reasoning and runtime discovery. It contains 263 paired tasks across 42 executable sandbox environments, where each pair consists of a should-act task and a should-abstain variant produced through a controlled perturbation to the instruction, tool, or environment state. To scale this paired design and resist data contamination, we propose AbstainGen, a fully automated pipeline that synthesizes sandbox environments and generates paired tasks end-to-end, validated by deterministic replay and semantic LLM judges; fresh task instances can be regenerated on demand, and three independent annotators rate 94-98% of sampled tasks as well-designed. Across 17 frontier LLMs in 4 agent harnesses, the best agent (Gemini 3.1 Pro) achieves only 59.5% paired accuracy (correct on both the act and abstain sides of each paired task). More importantly, abstention capability is largely independent of general task-solving capability, indicating that scaling task-solving alone will not close this gap. We further identify failure modes such as post-hoc abstention, in which agents execute irreversible actions before recognizing abstention triggers. Our code and dataset are open-sourced at agentabstain.github.io.