RLHF Fine-Tuning Quiz
5 questions Pass: 70% +25 pts (50 with Pro)
Quiz covering Advanced Training
What you'll be asked
- What is the primary purpose of Reinforcement Learning from Human Feedback (RLHF) in training Large Language Models?
- In the RLHF pipeline, what is the role of the Reward Model?
- Which algorithm is most commonly used to update the policy model during the RLHF phase to prevent the model from deviating too far from the original supervised fine-tuned model?
- Why is a KL-divergence penalty typically included in the PPO objective function during RLHF?
- When training a Reward Model, why is it common practice to use pairwise comparisons (A is better than B) rather than absolute scalar ratings (1-5 stars)?
Answers and explanations come after you submit.
Course Content
- Data Cleaning Techniques
- Data Cleaning Techniques Quiz5q
- Feature Scaling Normalization
- Feature Scaling Normalization Quiz5q
- Encoding Techniques
- Encoding Techniques Quiz5q
- Glue DataBrew
- Glue DataBrew Quiz5q
- SageMaker Feature Store
- SageMaker Feature Store Quiz5q
- Ground Truth Labeling
- Ground Truth Labeling Quiz5q
- SageMaker Training Jobs
- SageMaker Training Jobs Quiz5q
- Hyperparameter Tuning
- Hyperparameter Tuning Quiz5q
- Distributed Training
- Distributed Training Quiz5q
- Fine-Tuning Models
- Fine-Tuning Models Quiz5q
- Regularization Techniques
- Regularization Techniques Quiz5q
- Model Registry
- Model Registry Quiz5q
- Ensemble Methods
- Ensemble Methods Quiz5q