Ambiguous Exam Question Semantic Disambiguation QA Evaluation Set

#semantic disambiguation #ambiguity detection #answerability assessment #exam question analysis #educational assessment #question quality evaluation

5000
0.3GB
JSON
CATL
MOBIUSI INC

Updated:2026-10-19

AI Analysis & Value Prop

This evaluation set contains web text from digital question banks, focusing on issues that can affect question interpretation, including unclear references, polysemous terms, missing conditions, omitted expressions, and distracting wording. Each record is organized around question interpretation, ambiguity identification, answerability judgment, a reference answer, and supporting rationale, enabling systematic evaluation of how models handle semantic ambiguity in exam questions. It is suitable for digital question bank quality analysis, educational QA evaluation, and semantic understanding research.

Dataset Insights

Sample Examples

{}

Technical Specifications

Field	Type	Description
file_name	string	File name
answer_basis	string	An explanation of how the reference answer follows from the question conditions, interpretation, and required assumptions, identifying any ambiguity that could affect the conclusion.
webpage_text	string	Original webpage text containing a digital exam question and its context, preserving references, terminology, conditions, omissions, and distracting wording.
answerability	boolean	Whether the information provided in the question alone supports a unique, clear, and verifiable answer; true indicates that it does, and false indicates that it does not.
ambiguity_types	array	All major semantic ambiguity types present in the question, selected from predefined categories.
reference_answer	string	An answer, answer range, or conclusion that cannot be determined, based on the original question and any explicitly stated assumptions; key intermediate results may be included.
ambiguity_evidence	array	Words, phrases, sentences, or contextual clues quoted or summarized from the question to support the ambiguity judgment.
answerability_reason	string	An explanation of why the question is answerable or unanswerable, focusing on information sufficiency, answer uniqueness, and the effect of ambiguity on solving it.
required_assumptions	array	Prerequisites, definitions, values, or contextual information that must be added to provide a definite interpretation or answer; enter no additional assumptions if none are needed.
clarification_question	string	A question that would supply missing conditions or resolve multiple interpretations when ambiguity affects answering; enter no clarification needed when the question is clear.
question_interpretation	string	A description of the intended subject, question objective, and calculation or reasoning relationships, including how ambiguous content is interpreted.

Compliance Statement

Authorization Type	Proprietary - Commercial AI Training License (No Redistribution)
Commercial Use	Requires exclusive subscription or authorization contract (monthly or per-invocation charging)
Privacy and Anonymization	No PII, no real company names, simulated scenarios follow industry standards
Compliance System	Compliant with China's Data Security Law / EU GDPR / supports enterprise data access logs

Frequently Asked Questions

What does the dataset contain?: The dataset includes web text from digital question banks, along with annotations for question interpretation, ambiguity types and evidence, answerability, required assumptions, reference answers, and answer rationales. These annotations support analysis of semantic ambiguity and its impact on answering.

Which capabilities can this dataset evaluate?: It can evaluate a model's understanding of exam questions, including its ability to identify unclear references, polysemous terms, missing conditions, omitted expressions, and distracting wording; judge whether the available information supports a definite answer; and explain its judgment with textual evidence.

What are the dataset size and quality characteristics?: The available information does not specify the number of samples, detailed coverage, or quality review process, so the dataset size should not be inferred. Consult the product details for sample statistics, annotation guidelines, and quality information before purchase or use.

How can I obtain the dataset, and what should I check before using it?: The acquisition method, file format, and licensing terms are determined by the provider's product page or agreement. Before use, verify the permitted scope and whether commercial use is allowed. When evaluating models, account for the stated assumptions and answer rationales rather than treating every ambiguous question as a standard-answer error.

Can't find the data you need?

Post a request and let data providers reach out to you.

Get this Dataset

Verified for Enterprise Use

Cite this Work

@dataset{Mobiusi2026,
  title={Ambiguous Exam Question Semantic Disambiguation QA Evaluation Set},
  author={MOBIUSI INC},
  year={2026},
  url={https://www.mobiusi.com/datasets/7e8e066c1ffd7394f39612559823b4fd},
  urldate={2026-10-15},
  keywords={ambiguous exam question benchmark,question semantics evaluation,question bank QA data,question ambiguity detection,answerability assessment,educational QA benchmark,answer rationale analysis,digital question bank text},
  version={1.0}
}

Using this in research? Please cite us.

placeholder

Products

Scene

Domain

Modality

Task