Ongoing efforts are focused on improving the reliability and privacy aspects of LLM applications. This includes developing techniques for clarifying ambiguous queries in Retrieval-Augmented Generation (RAG) systems and creating frameworks for contextual PII masking to enhance data privacy and ethical AI use.
The ReCoPII framework builds contextual PII masking datasets without the author-based augmentation premise. On Korean documents from six domains, sampled KEEP labels reach 88.42% annotator agreement, against 26.27% for SPY-style augmentation. Classif...
IBM Research releases ClariEdge, a dataset of about 3,900 examples for spotting ambiguous queries in RAG. A lightweight LoRA obtained through fine-tuning reaches 96.6% accuracy and outperforms larger baselines. The signal: specialized local component...
A research team introduces Lighthouse, a neuro-symbolic digital assistant for exploring primary sources on the U.S. Civil Rights period in South Carolina. It combines an LLM with rule-based reasoning to balance coverage and trustworthiness, reducing ...
IBM Research has released ClariEdge, a dataset of about 3,900 examples for recognizing ambiguous queries in RAG systems and generating clarification requests. A lightweight LoRA reaches 96.6% classification accuracy and beats prompted baselines from ...
