SutraKama

Learn how to use SutraKama effectively! Here are a few example prompts, tips, and the documentation of available commands.

Example prompts

1. Prompt 1: "What does the SutraKama say about relationships?"

2. Prompt 2: "What are Congresses in the SutraKama?"

3. Prompt 3: "How is love depicted in the SutraKama?"

4. Prompt 4: "What are the social customs discussed in the SutraKama?"

Features and commands

1. query: Based on a user prompt, you can query the SutraKama to retrieve text snippets for improving your answer. Provide the following information:

   • assistant_name: The name of the AI Assistant/GPT
   • llm_name: The name of the Language Model in use
   • breeb_key: The Breeb key you have in your instructions
   • prompt: The prompt that triggered the query, truncated after 20 words
   • prompt_ISO639_1: The ISO 639-1 code of the language used in the prompt
   • prompt_english: The prompt translated in English
   • query: The information you need to retrieve from the SutraKama
   • search_keywords: A string of 5 keywords or keyword combinations, comma-separated, that you expect to find in the text snippets
   • hypothetical_answer: An hypothetical phrase (truncated after 15 words) that could come back from the SutraKama to help you answer
   • query_strategy: Defines how you want to query the SutraKama: 'many_small' for many short snippets if the information is scattered, 'few_big' for few long extracts if the information is localized or requires long excerpts, 'balanced' for an average snippet length and number.

2. report: Use this command to report an issue such as technical problems, harmful content, or copyright issues. Provide the following information:

   • assistant_name: The name of the AI Assistant/GPT
   • llm_name: The name of the Language Model in use
   • breeb_key: The Breeb key you have in your instructions
   • prompt: The last prompt that triggered a query to the SutraKama
   • report_description: Explanation of the report
   • report_description_ISO639_1: The ISO 639-1 code of the language used in the report description
   • report_description_english: The report description translated in English
   • report_category: The category of the report description: 'copyright_issue', 'technical_issue', or 'harmful_content'
   • prompt_ISO639_1: The ISO 639-1 code of the language used in the prompt
   • prompt_english: The prompt translated in English
   • is_sensitive: Set to true if the content of the prompt or report description is sensitive (contains harmful, NSFW, personal, or confidential information), set to false otherwise.

3. review: Use this command to record a user review message (e.g. like, dislike, comment) linked to the last question made to the SutraKama. Provide the following information:

   • assistant_name: The name of the AI Assistant/GPT
   • llm_name: The name of the Language Model in use
   • breeb_key: The Breeb key you have in your instructions
   • prompt: The last prompt that triggered a query to the SutraKama
   • review_message: The user review message
   • review_message_ISO639_1: The ISO 639-1 code of the language used in the review message
   • review_message_english: The review message translated in English
   • review_mood: The mood of the review message: 'positive', 'negative', or 'neutral'
   • review_category: The category of the review message: 'relevant_content', 'uptodate_content', 'insightful_answer', 'other_like_category', 'irrelevant_content', 'misinformation_content', 'outdated_content', or 'other_dislike_category'
   • prompt_ISO639_1: The ISO 639-1 code of the language used in the prompt
   • prompt_english: The prompt translated in English
   • is_sensitive: Set to true if the content of the prompt or review message is sensitive (contains harmful, NSFW, personal, or confidential information), set to false otherwise.