Legal contract understanding often requires connecting information distributed across multiple clauses and cross-referenced provisions. While Retrieval-Augmented Generation (RAG) grounds Large Language Models (LLMs) in relevant evidence, conventional retrieval primarily prioritizes direct query relevance and may overlook supporting contractual context. We propose the Multi-View Contract Graph (MCG), a contract-specific knowledge graph that models semantic, entity, keyword, and structural relationships among clauses. By traversing these relationships, MCG retrieves supporting evidence that may not directly match the query but is necessary for accurate interpretation. We evaluate MCG on the Contract Understanding Atticus Dataset (CUAD), covering 15 information categories, including audit rights, governing laws, and liability-related provisions. Across five open-source LLMs, MCG consistently outperforms traditional retrieval, achieving a best F₁ score of 0.460 with Mistral-7B-Instruct-v0.3. Compared with BM25-RAG and ColBERT-RAG, the best MCG configuration improves F₁ by approximately 10.0% and 12.2%, respectively. These results demonstrate that modeling clause-level relationships improves retrieval and information extraction for legal contracts.

Policy documents often employ intricate and technical language, posing comprehension challenges for policyholders and increasing the risk of misinterpretation, financial losses, and legal disputes. To address these issues, we propose an automated framework leveraging Retrieval-Augmented Generation to identify and clarify potentially mis-interpretable paragraphs within policy documents. The framework consists of two key modules: the Annotation module and the Rectification module. The Annotation module employs both paragraph-level and document-level contextual reasoning to classify paragraphs into categories indicative of potential misinterpretation. The Rectification module resolves these ambiguities by generating targeted interpretation queries, retrieving relevant document-level context, and incorporating external knowledge sources. Applied to a corpus of 240 real-world policy documents, the Annotation module produced a benchmark dataset comprising 11,000 annotated paragraphs, enabling systematic evaluation of interpretability issues. We assessed the dataset’s quality through expert-driven manual reviews and large-scale automated evaluations using fine-tuned Pretrained Language Model. For the Rectification module, we evaluated five open-source Large Language Models: Mistral-2-7B, Mistral-3-7B, LLaMA-2-7B, LLaMA-3-8B, andSaul-7B. Among these, Mistral-2-7B achieved the highest human evaluation scores: 0.912 for Clarity, 0.914 for Fidelity, and 0.934 for Usefulness. This work demonstrates the practical feasibility of utilizing automated frameworks to enhance the clarity and comprehensibility of complex policy documents, thereby mitigating risks associated with misinterpretation and its adverse consequences.

The rapid proliferation of mobile applications has resulted in a vast volume of user-generated reviews, which often contain critical insights into bugs, feature demands, and overall user satisfaction. Efficiently organizing and interpreting this feedback is essential for developers seeking to enhance app quality and user experience. In this study, we propose a fine-grained, multilabel classification schema for app reviews, structured as a two-level taxonomy. The first level categorizes reviews into three primary classes: Bug Report, Feature Request, and App Satisfaction Feedback. The second level further refines these categories into subtypes addressing specific concerns such as performance, usability and privacy. To automate this classification, we fine-tune several state-of-the-art large language models (LLMs), including LLaMA, Gemma, Mistral, Qwen, on a manually annotated dataset spanning 10 diverse application domains. Among these, LLaMA 3.1 achieved the highest performance, with weighted F1 scores of 0.90 and 0.79 at the first and second taxonomy levels, respectively. To assess scalability, we applied the fine-tuned LLaMA 3.1 model to label an additional 10,000 unlabeled reviews, creating a silver-standard corpus. This taxonomy-driven approach enables more nuanced analysis of user feedback, supporting targeted app improvements and fostering enhanced user satisfaction.

Enterprises frequently enter into commercial contracts that can serve as vital sources of project-specific requirements. Contractual clauses are obligatory, and the requirements derived from contracts can detail the downstream implementation activities that non-legal stakeholders, including requirement analysts, engineers, and delivery personnel, need to conduct. However, comprehending contracts is cognitively demanding and error-prone for such stakeholders due to the extensive use of Legalese and the inherent complexity of contract language. Furthermore, contracts often contain ambiguously worded clauses to ensure comprehensive coverage. In contrast, non-legal stakeholders require a detailed and unambiguous comprehension of contractual clauses to craft actionable requirements. In this work, we introduce a novel legal NLP task that involves generating clarification questions for contracts. These questions aim to identify contract ambiguities on a document level, thereby assisting non-legal stakeholders in obtaining the necessary details for eliciting requirements. This task is challenged by three core issues: (1) data availability, (2) the length and unstructured nature of contracts, and (3) the complexity of legal text. To address these issues, we propose ConRAP, a retrieval-augmented prompting framework for generating clarification questions to disambiguate contractual text. Experiments conducted on contracts sourced from the publicly available CUAD dataset show that ConRAP with ChatGPT can detect ambiguities with an F2 score of 0.87. 70% of the generated clarification questions are deemed useful by human evaluators.

With the growing number of mobile users, app development has become increasingly lucrative. Reviews on platforms such as Google Play and Apple App Store provide valuable insights to developers, highlighting bugs, suggesting new features, and offering feedback. However, many reviews contain typos, spelling errors, grammar mistakes, and complex sentences, hindering efficient interpretation and slowing down app improvement processes. To tackle this, we introduce RARE (Repository for App review REfinement), a benchmark dataset of 10,000 annotated pairs of original and refined reviews from 10 mobile applications. These reviews were collaboratively refined by humans and large language models (LLMs). We also conducted an evaluation of eight state-of-the-art LLMs for automated review refinement. The top-performing model (Flan-T5) was further used to refine an additional 10,000 reviews, contributing to RARE as a silver corpus.

Commercial contracts are known to be a valuable source for deriving project-specific requirements. However, contract negotiations mainly occur among the legal counsel of the parties involved. The participation of non-legal stakeholders, including requirement analysts, engineers, and solution architects, whose primary responsibility lies in ensuring the seamless implementation of contractual terms, is often indirect and inadequate. Consequently, a significant number of sentences in contractual clauses, though legally accurate, can appear unfair from an implementation perspective to non-legal stakeholders. This perception poses a problem since requirements indicated in the clauses are obligatory and can involve punitive measures and penalties if not implemented as committed in the contract. Therefore, the identification of potentially unfair clauses in contracts becomes crucial. In this work, we conduct an empirical study to analyze the perspectives of different stakeholders regarding contractual fairness. We then investigate the ability of Pre-trained Language Models (PLMs) to identify unfairness in contractual sentences by comparing chain of thought prompting and semi-supervised fine-tuning approaches. Using BERT-based fine-tuning, we achieved an accuracy of 84% on a dataset consisting of proprietary contracts. It outperformed chain of thought prompting using Vicuna-13B by a margin of 9%.