GRAID: Synthetic data generation with geometric constraints and multi-agentic reflection for harmful content detection
A novel pipeline that leverages Large Language Models (LLMs) for dataset augmentation.
We address the problem of data scarcity in harmful text classification for guardrailing applications and introduce GRAID (Geometric and Reflective AI-Driven Data Augmentation), a novel pipeline that leverages Large Language Models (LLMs) for dataset augmentation. GRAID consists of two stages: (i) generation of geometrically controlled examples using a constrained LLM, and (ii) augmentation through a multi-agentic reflective process that promotes stylistic diversity and uncovers edge cases. This combination enables both reliable coverage of the input space and nuanced exploration of harmful content. Using two benchmark data sets, we demonstrate that augmenting a harmful text classification dataset with GRAID leads to significant improvements in downstream guardrail model performance.
Latest publications
RECAP: Regression evaluation for continual adaptation of prompts
A benchmark that measures continual-learning phenomena at the constraint level for prompt-level adaptation methods.
EMNLPHarmonizing diverse models: a layer-wise merging strategy
An approach combining systematic synthetic data generation, triplet loss for embeddings, and layer-wise model merging.
EMNLPSPARC-RAG: Adaptive Sequential–Parallel Scaling with Context Management for Retrieval-Augmented Generation
A multi-agent framework that coordinates sequential and parallel inference-time scaling.
EMNLP