Qworld: Question-Specific Evaluation Criteria for LLMs

TL;DR

Qworld introduces a hierarchical, question-specific evaluation method for LLMs that generates detailed criteria tailored to each question, improving the assessment of nuanced capabilities.

Contribution

It presents a novel recursive expansion tree approach to generate tailored evaluation criteria for individual questions, enhancing the granularity and relevance of LLM assessments.

Findings

Covers 89% of expert criteria on HealthBench

Generates 79% novel, validated criteria

Reveals capability differences across models in nuanced dimensions

Abstract

Evaluating large language models (LLMs) on open-ended questions is difficult because response quality depends on the question's context. Binary scores and static rubrics fail to capture these context-dependent requirements. Existing methods define criteria at the dataset level or generate them in a single pass, which limits their ability to explore the evaluation space implied by each question. We introduce One-Question-One-World (Qworld), a method that generates question-specific evaluation criteria using a recursive expansion tree. Given a question, Qworld decomposes it into scenarios, perspectives, and fine-grained binary criteria through structured hierarchical and horizontal expansion. The resulting criteria specify what a high-quality answer must address for that question. On HealthBench, Qworld covers 89% of expert-authored criteria and generates 79% novel criteria validated by human experts. Experts rate Qworld criteria higher in insight and granularity than those produced by prior methods. When applied to 11 frontier LLMs on HealthBench and Humanity's Last Exam, Qworld reveals capability differences in dimensions such as long-term impact, equity, error handling, and interdisciplinary reasoning that coarse rubrics do not distinguish. By formulating criteria generation as structured coverage of question-implied evaluation axes, Qworld enables evaluation that adapts to each question rather than relying on fixed task-level criteria.