Knowledge base name: knowledge-base-ai-email-auto-response
IAM permissions: Create and use a new service role
Data source: Amazon S3
Choose Next
At this step, i choose Titan Embeddings G1 - Text v1.2 > Next
The new Titan Embeddings G1 – Text v1.2 can intake up to 8k tokens and outputs a vector of 1536 dimensions. The model also works in 25+ different languages. The model is optimized for text retrieval tasks but can also perform additional tasks such as semantic similarity and clustering. Titan Embeddings G1 – Text v1.2 also supports long documents, however, for retrieval tasks it is recommended to segment documents into logical segments (such as paragraphs or sections).
The final step is choose Create knowledge base
The knowledge base is create successfully, you must remember the knowledge base id for setting up our lambda function, then we need to sync them before use it.
Choose our faqs.txt document from Data source > Sync > Select button Select model at the right panel > Claude 3 Haiku. Then you can ask the knowledge base about our policies in the faqs document.
As you can see, they can read the document and response them.
This file is a sample document about policies of a cat shop i generated it using Gemini. This will include some sections:
About Our Store
Additional Questions
Return Policy
Protecting Your Personal Information
As you can see, this document is very humanity with format is txt, you can also using pdf, docx,… with data is your personal documents. You can read the FAQs about Amazon Bedrock here