Activity (Process)

Workflow for Crowdsourcing Fine-Tuning Data

A typical workflow for crowdsourcing fine-tuning data begins with allowing users to submit a wide range of questions. Subsequently, responses are generated, either manually by humans or automatically by an LLM. The final stage involves manual annotation and correction of these responses to ensure data quality.

0

1

Updated 2025-10-10

Contributors are:

Who are from:

Tags

Ch.4 Alignment - Foundations of Large Language Models

Foundations of Large Language Models

Foundations of Large Language Models Course

Computing Sciences