Book

nlp

Pocket Reference Title

RLHF

Proposed Content

Pocket Ref for human alignment with RLHF

Rationale

One of the most popular alignment methods for LLMs

Content Types

• Theoretical foundations
• Mathematical formulations
• Code examples
• Diagrams/visualizations
• Practical applications
• Common pitfalls/challenges

Additional Resources

• https://magazine.sebastianraschka.com/p/llm-training-rlhf-and-its-alternatives
• https://magazine.sebastianraschka.com/p/tips-for-llm-pretraining-and-evaluating-rms?open=false#%C2%A7rlhf-vs-direct-preference-optimization-dpo
• InstructLLM (https://arxiv.org/pdf/2203.02155)