Learn Before
Example of a User Prompt in RLHF
An example of a user prompt that an LLM might receive at the beginning of the Reinforcement Learning from Human Feedback (RLHF) process is: 'How can I live a more environmentally friendly life?' This type of input is used to generate multiple responses from the model, which are then evaluated by humans.
0
1
Tags
Ch.2 Generative Models - Foundations of Large Language Models
Foundations of Large Language Models
Foundations of Large Language Models Course
Computing Sciences
Related
Training a Reward Model with Preference Data
Techniques for Generating Diverse Outputs in RLHF
A team is developing a system to align a language model with human preferences. Their data collection process involves providing a prompt to an existing, fine-tuned model, which then generates a single response. A human labeler then assigns a quality score from 1 to 10 to this single response. This process is repeated for thousands of different prompts. What is the most significant flaw in this methodology for the purpose of creating a robust preference-based reward model?
Arrange the following steps in the correct chronological order to describe the data collection process for training a reward model.
Designing a Data Collection Pipeline for a Creative Writing Assistant
Learn After
Imagine you are part of a team training a new AI assistant. A key step in this process involves providing the AI with a single question, generating multiple different responses to it, and then having human reviewers rank these responses from best to worst. This helps the AI learn what constitutes a high-quality answer. Which of the following questions would be most effective for this specific training step?
Optimizing an AI for Creative Brainstorming
Evaluating Prompts for AI Training
Learn After
Imagine you are part of a team training a new AI assistant. A key step in this process involves providing the AI with a single question, generating multiple different responses to it, and then having human reviewers rank these responses from best to worst. This helps the AI learn what constitutes a high-quality answer. Which of the following questions would be most effective for this specific training step?
Optimizing an AI for Creative Brainstorming
Evaluating Prompts for AI Training
Related
Training a Reward Model with Preference Data
Techniques for Generating Diverse Outputs in RLHF
A team is developing a system to align a language model with human preferences. Their data collection process involves providing a prompt to an existing, fine-tuned model, which then generates a single response. A human labeler then assigns a quality score from 1 to 10 to this single response. This process is repeated for thousands of different prompts. What is the most significant flaw in this methodology for the purpose of creating a robust preference-based reward model?
Arrange the following steps in the correct chronological order to describe the data collection process for training a reward model.
Designing a Data Collection Pipeline for a Creative Writing Assistant