Employment Agreement Conflicting Clause Retrieval Benchmark Dataset
#legal evidence retrieval #clause retrieval #conflicting-clause context identification #employment agreement review #conflicting clause analysis #labor contract evidence retrieval
- 5000
- 0.3GB
- JSON
- CATL
MOBIUSI INC
Updated:2026-10-18
AI Analysis & Value Prop
This benchmark covers cases involving conflicting language, overlapping scope, or unclear precedence within an employment agreement, between a main agreement and its attachments, or across document versions. Each case includes contract content, version information, and a retrieval question, along with annotations for conflict assessment and evidence fragments that should be retrieved together. It supports evaluation of whether legal RAG systems retrieve the full conflict context rather than returning only a single similar clause.
Dataset Insights
Sample Examples
{}
Technical Specifications
| Field | Type | Description |
|---|---|---|
| file_name | string | File name |
| case_source | object | The original input object containing contract document content, version information, and the retrieval question. |
| retrieval_target | object | The target requirements used to evaluate whether a retrieval system returns the complete conflict context. |
| evaluation_labels | object | Standard annotations describing candidate evidence relevance, necessity, and role in the conflict. |
| evidence_fragments | array | Related contract clauses or contextual fragments that should be jointly covered by the retrieval results. |
| conflict_assessment | object | The conflict type, scope, and precedence assessment derived from the source contract documents and retrieval question. |
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?
- Cases cover conflicting language, overlapping scope, or unclear precedence within an employment agreement, between a main agreement and its attachments, and across document versions. The records include case inputs, conflict assessments, evidence fragments to retrieve together, and evaluation annotations.
- How does the dataset evaluate retrieval quality?
- It focuses on whether retrieval results jointly cover the clauses and context needed to understand a conflict. Retrieval coverage targets and annotations for candidate evidence relevance, necessity, and conflict role are provided. Evaluation metrics and workflows can be configured for the retrieval system being tested.
- How can I obtain the dataset?
- Check the dataset product page for current access options, delivery formats, and available versions. If these details are not listed, contact the data provider for confirmation.
- What are the intended uses, and how can I confirm licensing?
- The dataset is suitable for legal RAG retrieval evaluation, employment agreement review assistance, contract version comparison, and legal evidence retrieval research. Commercial use, redistribution, and other licensing permissions are governed by the terms on the product page or those provided by the data provider.
Can't find the data you need?
Post a request and let data providers reach out to you.
Get this Dataset
Verified for Enterprise UseCite this Work
@dataset{Mobiusi2026,
title={Employment Agreement Conflicting Clause Retrieval Benchmark Dataset},
author={MOBIUSI INC},
year={2026},
url={https://www.mobiusi.com/datasets/38d1381f77bb613bb72c5583da01f9f0?dataset_scene_id=8},
urldate={2026-10-16},
keywords={employment agreement dataset,contract conflict retrieval,RAG retrieval benchmark,legal document data,employment contract review,conflicting clause detection,evidence fragment retrieval,legal RAG evaluation},
version={1.0}
}Using this in research? Please cite us.
placeholder
placeholder
placeholder
placeholder
placeholder
placeholder
placeholder