Named after Kahnemans idea of the intuitive thinking system in human cognition.
Use cases
- Spam detection
- Content moderation
- Relevance classification
- SEO (rearranging links)
- Sorting text into folders, projects etc.
- Reranking for Retrieval augumented generation (RAG)
- LLM Routing
- Recommendations
Jev
Jev is a proprietary AI model for this usecase.
https://typesafe.ai/
Integration for LangChain: https://docs.langchain.com/oss/python/integrations/providers/typesafe
OpenRouter Integration: https://openrouter.ai/~typesafe/jev-latest
Primitives
- Choice: One Option from a given set
- Score: Rate a statement against given ordered levels e. g. (high, medium, low)
- Noul: Evaluate yes/no with propability 0 - 1
Other models
Resources
- Layla, a system one model which is open-source: https://laya.convaiinnovations.com/
- Guide for using Jev for classification: https://openrouter.ai/docs/cookbook/evaluate-and-optimize/jev-classification