Cleanlab for the Legal Industry
Case StudyDocument Review
Cleanlab Studio was used by a customer to identify errors in relevance annotations for documents during the e-discovery phase of a legal proceeding. Cleanlab automatically identified a vast number of important documents that paralegals accidentally failed to annotate as relevant. Relevance-prediction models trained inside Cleanlab Studio were 15% more accurate than the customer’s own models. Similar benefits were observed for dealing with annotations for privileged content.
HOW CLEANLAB IMPROVES ANALYSIS OF LEGAL DATA
Cleanlab Studio auto-corrects raw data to ensure reliable insights so you can make good legal decisions.
Model training & deployment only requires a few clicks (no technical knowledge necessary).
Cleanlab models produced in this seamless manner are more accurate than fine-tuned OpenAI LLMs (the state-of-the-art for text prediction) when applied to predict legal outcomes from court case descriptions.
Details in article: Improving Legal Judgement Prediction with Cleanlab Studio