Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.q-uestionable.ai/llms.txt

Use this file to discover all available pages before exploring further.

RXP requires two components: embedding models to vectorize text and domain profiles to provide standard queries.

​
Embedding Models

RXP uses local sentence-transformers models via the optional [rxp] dependencies (sentence-transformers + chromadb). All models run locally — no API keys needed.

​
Registered Models

qai rxp list-models
IDHuggingFace NameDimensionsDescription
minilm-l6sentence-transformers/all-MiniLM-L6-v2384Default. Open WebUI default embedding model.
minilm-l12sentence-transformers/all-MiniLM-L12-v2384Higher quality MiniLM variant.
bge-smallBAAI/bge-small-en-v1.5384BGE small English v1.5.

​
Using a Specific Model

qai rxp validate --profile hr-policy --model minilm-l12

​
Ad-Hoc HuggingFace Models

Any HuggingFace sentence-transformers model can be passed directly — if it’s not in the registry, RXP creates an ad-hoc config: 
qai rxp validate --profile hr-policy --model BAAI/bge-m3
The model is downloaded automatically on first use.

​
Comparing All Registered Models

qai rxp validate --profile hr-policy --model all
This runs validation against every registered model and prints a comparison table showing retrieval rates and mean poison ranks.

​
Domain Profiles

Profiles provide pre-configured corpus documents, poison documents, and test queries for specific knowledge domains.

​
List Profiles

qai rxp list-profiles
Each profile shows: ID, name, query count, and corpus document count.

​
Built-in Profiles

hr-policy — Simulated corporate HR policy knowledge base. 8 queries covering remote work, time off, benefits, dress code, performance reviews, parental leave, expenses, and holiday dates. 
qai rxp validate --profile hr-policy

​
Custom Queries

Instead of a profile, provide queries directly with --query (repeatable): 
qai rxp validate \
  --corpus-dir ./my_knowledge_base \
  --poison-file ./adversarial_doc.txt \
  --query "What is the refund policy?" \
  --query "How do I contact support?"

​
Custom Corpus

Provide your own corpus as a directory of .txt files: 
qai rxp validate \
  --corpus-dir ./knowledge_base \
  --poison-file ./adversarial_doc.txt \
  --query "What are the security guidelines?"
Each .txt file in the directory becomes one corpus document. The poison file is a single document that you’re testing for retrieval rank.

​
Model Selection Guide

All three registered models are local and free — no API keys needed. Choose based on your needs:
Use CaseRecommended
Quick testingminilm-l6 (default, fastest)
Higher qualityminilm-l12
Cross-model comparison--model all
Specific research modelPass HuggingFace name directly