VERA: Validation and Enhancement for Retrieval Augmented systems

Large language models (LLMs) exhibit remarkable capabilities but often produce inaccurate responses, as they rely solely on their embedded knowledge. Retrieval-Augmented Generation (RAG) enhances LLMs by incorporating an external information retrieval system, supplying additional context along with the query to mitigate inaccuracies for a particular context. However, accuracy issues still remain, as the model may rely on irrelevant documents or extrapolate incorrectly from its training knowledge. To assess and improve the performance of both the retrieval system and the LLM in a RAG framework, we propose \textbf{VERA} (\textbf{V}alidation and \textbf{E}nhancement for \textbf{R}etrieval \textbf{A}ugmented systems), a system designed to: 1) Evaluate and enhance the retrieved context before response generation, and 2) Evaluate and refine the LLM-generated response to ensure precision and minimize errors. VERA employs an evaluator-cum-enhancer LLM that first checks if external retrieval is necessary, evaluates the relevance and redundancy of the retrieved context, and refines it to eliminate non-essential information. Post-response generation, VERA splits the response into atomic statements, assesses their relevance to the query, and ensures adherence to the context. Our experiments demonstrate VERA's remarkable efficacy not only in improving the performance of smaller open-source models, but also larger state-of-the art models. These enhancements underscore VERA's potential to produce accurate and relevant responses, advancing the state-of-the-art in retrieval-augmented language modeling. VERA's robust methodology, combining multiple evaluation and refinement steps, effectively mitigates hallucinations and improves retrieval and response processes, making it a valuable tool for applications demanding high accuracy and reliability in information generation. .

Paper

References (16)

112024. Adapting LLMs to Downstream ApplicationsLarge Language Models in Cybersecurity: Threats, Exposure and Mitigation
122023 Llama 2: Open Foundation and Fine-Tuned ChatModels

Scroll for more · 4 remaining

Similar papers

© 2026 NYSGPT2525 LLC