Dataset Overview

Complete composition and core characteristics of the contract review error annotation data.

Dimension	Description	Details
Total Data Volume	One-million-scale review records	Including original review + error type + amendment proposal
	30+ contract types	Sale/lease/service/lending/guarantee/labor/technology/trade secret/non-compete restrictions, etc.
Reviewer perspectives	A/B/both perspectives	Covers the review needs of different perspectives
Risk level	High/Medium/Low, Level 3	Approximately 25% high, 45% medium, and 30% low
Error Type	5 major error categories	Errors in application of law/omission of provisions/logical contradictions/formatting issues/weak protection of interests
Review of focus	Each clause contains 2-5 review concerns	Liability for breach, intellectual property, confidentiality, dispute resolution, delivery and acceptance, etc.
Data Format	JSON / JSONL / CSV	Supports batch export integration via API

Dataset fields and definitions

Each review record contains a complete set of multi-dimensional annotation fields.

Contract fundamentals

Contract number, contract name, contract type, signing date, Party A/Party B information (de-identified)

Review records

Review ID, original clause text, reviewer perspective (A/B/both), risk rating, review concerns

Error annotations

Error type (5 major categories), correctness, explanation of the cause of the error, cited article of law

Amendment recommendations

Full text of the amended review, reasons for the amendment, alternative clause recommendations

Legal basis

Index of relevant laws and regulations, references to judicial interpretations, references to typical cases

Contract clause classification

Type of provision (definitions / rights and obligations / default provisions / dispute resolution, etc.), article numbering scheme

Data Sample

The JSON format presents a complete record of contract review errors.

{ "doc_id": "sample000042", "contract_name": "XX[]", "contract_type": "[]", "contract_date": "2024-03-15", "party_a": "XX[]Company（[de-identified]）", "party_b": "XX[]Company（[de-identified]）", "review_position": "[]", "clause_items": [ { "clause_id": "C001", "clause_title": "Law[]", "original_text": "1.1 《 and Law》Laws and Regulations...", "review_id": "R001", "risk_level": "[]", "review_focus": "Law[]", "is_correct": false, "error_type": "Law[]", "error_reason": "《Law》2021January1，Law", "revised_review": "1.1 《 and Law》Laws and Regulations...", "legal_basis": ["[]Law[]463[]", "Article 1260 of the Civil Code."] }, { "clause_id": "C002", "clause_title": "[]", "original_text": "8.2 []30%...", "review_id": "R002", "risk_level": "[]", "review_focus": "[]", "is_correct": false, "error_type": "[]", "error_reason": "，", "revised_review": "8.2 30%，...", "legal_basis": ["[]Law[]585[]", "Civil Code contract[]Law[]65[]"] } ] }

AI Application Scenarios

Intelligent Contract Review

Automatically identify legal risks and logical issues in contract terms and provide amendment recommendations

Review Quality Assessment

Assesses the accuracy and completeness of manual review opinions to support counsel quality control

Contractual Risk Early Warning

Risk scanning and early warning for new contracts based on historical error patterns

Automatic Clause Generation

Automatically generate standard compliance clause text based on contract type and review experience

Search and Discovery

Contract Review Dataset Legal AI Training Data Contract Term Mislabeling Intelligent Contract Review Legal Risk Identification Contract Review Misalignment LegalNLPCorpora Compliance Review Civil Code Contract Contract Text Analysis

Contract Review Error Annotation Dataset