Contract Clause Absence and Evidence Insufficiency Verification Benchmark

#clause absence verification #evidence sufficiency assessment #absence and uncertainty classification #contract clause verification #legal document review #contractual evidence sufficiency assessment
Records
3710
Size
0.4GB
Format
JSON
Provider
MOBIUSI INCMOBIUSI INC
Updated
v1.02026-11-04
AI Analysis & Value Prop

This benchmark focuses on negative claims about procurement and sales agreements. It provides contract text with known document coverage; the material may be complete or partial. Each example assesses whether the available material proves that a specified right, restriction, or responsibility is absent, distinguishing proven absence from insufficient evidence, contrary evidence, and inconclusive records. It supports training and evaluating legal document verifiers for contract review, evidence assessment, and cautious reasoning.

Overview

Dataset Insights

Preview

Sample Examples

sample.json
{}

Schema

Technical Specifications

file_namestringFile name
absence_claimstringNegative claim asserting that the contract does not contain a specified right, restriction, or responsibility.
contract_textstringText of the supplied procurement or sales contract, which may be complete or partial.
document_scopestringKnown coverage of the supplied contract material.
evidence_spansarrayRelevant quotations or text excerpts supporting the assessment; an empty array may be used when no relevant span is available.
absence_assessmentstringAnnotated outcome distinguishing proven absence from mere lack of evidence, contrary evidence, or an inconclusive record.
assessment_rationalestringExplanation of why the available contract material does or does not establish the claimed absence, taking document scope into account.

Trust

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

FAQ

Frequently Asked Questions

What does the benchmark contain?
Examples include procurement or sales contract text, a claim that a specified right, restriction, or responsibility is absent, the known scope of the material, and an assessment with its rationale. Relevant evidence excerpts may also be provided.
How are incomplete contract texts handled?
Assessments account for the known document scope. A clause is not treated as absent merely because it was not found in a supplied excerpt. Labels can distinguish proven absence, insufficient evidence, contrary evidence, and an inconclusive record.
Which tasks is this dataset suited for?
It supports training and evaluation of legal document verifiers, contract review, negative claim verification, and evidence sufficiency assessment. It can also evaluate whether models handle partial materials and uncertain conclusions cautiously.
How can I access the data, and what are the licensing terms?
Refer to the dataset product page or provider for current access options, available versions, and licensing terms. Check the applicable license before use and ensure compliance with relevant legal, privacy, and data security requirements.

Can't find the data you need?

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

Get this Dataset

High-quality data, ready for your next AI breakthrough.

Cite this Work

@dataset{Mobiusi2026,
  title={Contract Clause Absence and Evidence Insufficiency Verification Benchmark},
  author={MOBIUSI INC},
  year={2026},
  url={https://www.mobiusi.com/datasets/3c2447a5980214b84add7fe83bd1f757?dataset_scene_cate_type=2},
  urldate={2026-10-22},
  keywords={contract clause absence,contract evidence verification,purchase agreement review,sales contract analysis,negative claim verification,evidence sufficiency,legal document benchmark,AI contract review},
  version={1.0}
}

Using this in research? Please cite us.

placeholder
placeholder
placeholder
placeholder
placeholder
placeholder
placeholder

Popular Dataset Searches