Merge pull request 'Minor cleanup' (#7) from FEAT-minor-cleanup into master

Reviewed-on: #7

README.md

@@ -1,13 +1,15 @@
-# Citation Sentinel © 2023 @sjDev
+# Citation Sentinel - 2025 @sjDev - LICENSE: MIT
 
 
-![Citation Sentinel UI showing the source sidebar, chat answer with citations, and groundedness badge](demo.jpg)
+![Citation Sentinel UI showing the source sidebar, natural language answer with citations to the source, and groundedness badge](demo.jpg)
 
+Image one:  source-grounded research assistant for applied use of LLMs. 
 
-Image one:  source-grounded research assistant that empowers users to transform an otherwise-unmanageably-large corpus of data on any topic of interest into refined, easily-digested subtopics and pose focused inquiries, to instantly receive concise, quantifiably-evaluated, accurate responses.  The system also provides suggested follow-up questions further refining in user inquiries by detecting the query goals.
+**Citation Sentinel leverages multiple Large Language Models for analyzing and conveniently working with large data corpora, when query-result accuracy is of the highest value.  Examples include complex litigation (i.e. Patent/IP), technical medical data exploration (reviewing health records, test results, or studies to find hidden patterns, trends, and answers), advanced LLM development and refinement.**
 
+Citation Sentinel empowers users to transform an otherwise-unmanageably-large corpus of data on any topic of interest into refined, easily-digested subtopics and pose focused inquiries, to instantly receive concise, quantifiably-evaluated, accurate responses.  The system also provides suggested follow-up questions further refining in user inquiries by detecting the query goals.
 
-Image Two:  demo co clickable inline citations that take user to source-grounded research basis/knowledge corpus supporting query response, with rated “groundedness” score.
+ALSO NOTE IN DEMO: clickable inline citations (in blue) that take user to source-grounded research basis/knowledge corpus supporting query response, with rated “groundedness” score.
 
 
 # Reliability and Understandability
@@ -22,7 +24,7 @@ Groundedness metrics quantify how well an AI-generated answer is supported by re
 # Architecture - Basic Overview
 
 
-Built with a React/Vote frontend and a Typescript/Node/Express backend, using Anthropic Claude for generation, OpenAI Whisper for video audio track transcription, Voyage AI voyage-3 for embeddings and Voyage AI rerank-2 for response cosine similarity scoring, to generate the "groundedness" score, presented as an easily-understandable red/yellow/green "badge" style tooltip in the response (see above.)
+Built with a React/Vite frontend and a Typescript/Node/Express backend, using Anthropic Claude for NL response generation, OpenAI Whisper for video audio track transcription, Voyage AI voyage-3 for embeddings and Voyage AI rerank-2 for response cosine similarity scoring (to generate the "groundedness" score) presented as an easily-understandable red/yellow/green "badge" style tooltip in the response (see image above.)
 
 
 ## Query pipeline
@@ -30,6 +32,19 @@ Built with a React/Vote frontend and a Typescript/Node/Express backend, using An
 
 Source Ingestion -> Parsing -> Chunking -> Embedding (using Voyage AI voyage-3 model) -> Storage (vector store) -> { user query submission } -> Evaluation of User Query -> Retrieval -> Ranking -> Response Generation (Using Anthopic's claude-opus-4-6 model) -> Response Groundedness Scoring (using Voyage AI rerank-r model)
 
+## Why Database Math Can't Replace Groundedness Scoring
+
+Many recent additions to the Vector database/store product space use distance metrics (Cosine, L2, Inner Product) for Bi-Encoder Similarity. They compare the vector of the user query against the vectors of chunks independently.
+
+The rerank-r model at the end of the Source Sentinel pipeline performs Cross-Encoder Evaluation, taking two entirely separate text inputs:
+
+1) The generated response from claude-opus-4-6 and 
+
+2) The raw source chunks
+
+It processes them simultaneously through deep attention layers to check for hallucinations, missing context, and factual alignment.
+
+A vector database cosine metric only calculates how close two embeddings are in coordinate space. It **does not read the generated text of a response to verify if it accurately reflects the source chunks**.
 
 ## Prerequisites
 

server/src/services/documentService.ts

@@ -104,11 +104,11 @@ export async function generateStudyGuide(sourceGroups: SourceGroup[]): Promise<S
   const sources = buildSourceBlock(sourceGroups);
   const start = Date.now();
 
-  const prompt = `You are an expert educator. Given the source documents below, produce a comprehensive study guide. Follow these rules:
+  const prompt = `You are an expert reasearch assistant. Given the source documents below, produce a comprehensive guide. Follow these rules:
 
 1. Create a structured outline of the main ideas organized into logical sections.
 2. For each section, provide: bullet-point key concepts, key terms with definitions, and 2-3 self-review questions.
-3. At the end, provide mnemonic devices or simplified restatements to help memorize the hardest concepts.
+3. At the end, provide mnemonic devices or simplified restatements of the hardest concepts.
 4. Be concise but thorough. Use simple, clear language.
 
 --- SOURCE DOCUMENTS ---