Translations:FACTS About Building Retrieval Augmented Generation-based Chatbots/65/pt: Difference between revisions

Latest revision as of 07:30, 20 February 2025

Information about message (contribute)

This message has no documentation. If you know where or how this message is used, you can help other translators by adding documentation to this message.

Message definition (FACTS About Building Retrieval Augmented Generation-based Chatbots)

'''Enterprise Guardrailing''': Implementing guardrails that align generative AI responses with specific enterprise policies and rules is essential. These guardrails help mitigate risks by ensuring that Chatbot-generated content adheres to established norms and ethical guidelines, preventing potential legal and reputational damage. In NVInfo bot, we implemented many guardrails in LLM prompts initially. However, later realized that not all LLMs follow these prompts consistently. Therefore, we implemented these guardrails during pre and post-processing of queries and responses respectively using Nemo Guardrails ([[#bib.bib13|13]]).

Proteção Empresarial: Implementar proteções que alinhem as respostas de IA generativa com políticas e regras específicas da empresa é essencial. Essas proteções ajudam a mitigar riscos, garantindo que o conteúdo gerado pelo Chatbot adira a normas estabelecidas e diretrizes éticas, prevenindo possíveis danos legais e de reputação. No bot NVInfo, implementamos muitas proteções em prompts de LLM inicialmente. No entanto, percebemos mais tarde que nem todos os LLMs seguem esses prompts de forma consistente. Portanto, implementamos essas proteções durante o pré e pós-processamento de consultas e respostas, respectivamente, usando Nemo Guardrails (13).