RAG
RAG & Knowledge Retrieval
Retrieval-Augmented Generation, knowledge conflicts, and the Input-Regime Audit framework for production RAG systems.
FAQ
What is the Input-Regime Audit framework for RAG?
The Input-Regime Audit is a framework for characterising the conflict patterns that cause RAG systems to fail. It identifies the input regimes (inter-context conflict, compliance regime, metadata weighting) that drive failures, and provides a 5-step diagnostic. Submitted to VecDB@VLDB 2026. arXiv:2606.27396.
What is Navigating the Knowledge Sea?
A paper on planet-scale answer retrieval using LLMs. We investigate methods for efficient answer retrieval across vast knowledge bases, addressing accuracy, scalability, and computational efficiency. Published February 2024 on arXiv (2402.05318). 6 citations as of 2026.
What is the CARS score and why does it hide ceiling effects?
CARS (Context-Adherence Rating Score) is the standard metric for RAG faithfulness. We show in the Input-Regime Audit paper that CARS hides the reasoning behind the rating and has ceiling effects that prevent differentiation between good and great RAG systems.
Related
- dipankar.org — strategic studies + advisor practice (broad public persona)
- dipankar.name — engineering work + AI agent infrastructure (canonical hub)
- dipankar.co — fractional CTO + consulting practice