Get Started
Home
Topics
Search
Library
Research questionHow can LLMs distinguish ambiguous inputs from gaps in their knowledge when estimating uncertainty?An input may support several plausible interpretations, making uncertainty appear to reflect missing model knowledge. Separating these sources is difficult when uncertainty estimates depend on generating answers for each interpretation, which can add cost and allow model-knowledge effects to distort the estimate.
AI
Evaluation & Benchmarks
Machine Learning
Natural Language Processing
Latest papersRecent research connected to this question, newest first.From Answers to Interpretations: Rethinking Ambiguity-Induced Aleatoric Uncertainty Estimation in LLMsThe source concerns LLMs handling ambiguous or underspecified inputs and evaluates ambiguity detection across three benchmarks. It provides theoretical analysis and evidence for estimating uncertainty from plausible interpretations without answering clarified inputs, with reported changes in detection performance, output-token cost, API-call cost, and correlation with epistemic uncertainty.research paper · Sep 3, 2026
Related questions
How should multilingual LLMs estimate uncertainty and calibrate abstention across languages and model sizes?When does an LLM’s verbal confidence reliably reflect its underlying uncertainty?How can we reliably assess whether conversational LLMs clarify ambiguity and recover user intent?How can LLMs produce reliable confidence estimates for deciding when to defer outputs to humans?