Uncertainty Decomposition for Clarification Seeking in LLM Agents
Quick Answer
This paper introduces a prompt-based uncertainty decomposition for LLM agents, enhancing clarification seeking capabilities.
Quick Take
The proposed method improves clarification F1 scores by 73% over ReAct+UE and 36% over UAM across five , including GPT-5.1 and GLM-4.7, on new benchmarks designed for underspecified tasks.
Key Points
- Introduces a prompt-based method for decomposing uncertainty in LLM agents.
- Enhances proactive clarification seeking in ambiguous task specifications.
- Achieves a 73% improvement in F1 scores on ALFWorld-Clarification.
- Evaluated across five LLM backbones including GPT-5.1 and GLM-4.7.
- New benchmarks include WebShop-Clarification and ALFWorld-Clarification.
Paper Resources
📖 Reader Mode
~2 min readAbstract:Recent position papers argue that the classical aleatoric/epistemic uncertainty framework is insufficient for interactive large language model (LLM) agents and call for underspecification-aware, decomposed, and communicable uncertainty representations that can unlock new agent capabilities such as proactive clarification seeking and shared mental-model building. Practical deployment constraints -- black-box APIs, interactive latency budgets, and the absence of labeled trajectories -- rule out logprob-based, multi-sampling, and training-based methods, leaving prompt-based estimation as the most viable family for surfacing such signals at deployment time. We answer this call with a simple prompt-based decomposition that separates action confidence from request uncertainty (u), enabling the agent to ask for clarification when the task specification is ambiguous. To evaluate it, we introduce two clarification-augmented benchmarks (WebShop-Clarification and ALFWorld-Clarification) in which 50% of tasks are deliberately underspecified, and systematically compare the proposed decomposition against ReAct+UE and Uncertainty-Aware Memory (UAM) across five LLM backbones (GPT-5.1, DeepSeek-v3.2-exp, GLM-4.7, Qwen3.5-35B, GPT-OSS-120B) on these variants together with the standard WebShop, ALFWorld, and REAL benchmarks for fault detection. Averaged across the five backbones, the proposed decomposition improves clarification F1 on ALFWorld-Clarification by 73% over ReAct+UE and by 36% over UAM, and leads clarification F1 on every backbone on WebShop-Clarification and on four of five backbones on ALFWorld-Clarification, indicating that the gains generalize beyond a single LLM.
| Comments: | 26 pages, 8 figures. Source code: this https URL |
| Subjects: | Artificial Intelligence (cs.AI); Computation and Language (cs.CL) |
| Cite as: | arXiv:2606.19559 [cs.AI] |
| (or arXiv:2606.19559v1 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2606.19559 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Gregory Matsnev [view email]
[v1]
Wed, 17 Jun 2026 19:59:32 UTC (1,420 KB)
— Originally published at arxiv.org
Want this in your inbox every morning?
Daily brief at your local 8am — bilingual EN/中文, free.
More from arXiv cs.AI
See more →HOBA: Hierarchical On-Policy Bidding Agents for Adaptive Online Advertising
HOBA (Hierarchical On-policy Bidding Agents) is a novel hierarchical reinforcement learning framework that enhances online advertising bidding systems by improving adaptability and reducing hyperparameter tuning costs. It utilizes a for hyperparameter inference, a SARSA agent for expert model selection, and a dynamic expert pool for bid execution, achieving a +3.6% increase in target cost during large-scale deployment and outperforming state-of-the-art baselines on AuctionNet.