Learn Before
Post-Pre-training Alignment Steps
After a Large Language Model undergoes pre-training on large volumes of unlabeled data, a typical subsequent phase involves a two-step alignment process to ensure its behavior is consistent with human intentions.
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
Guidance Sources for LLM Alignment
Desirable Attributes of Aligned LLMs
Aligning Large Language Models with Human Values
Challenges in LLM Alignment
Increased Research in LLM Alignment due to Control Concerns
Instruction Alignment
Necessity of Multiple LLM Alignment Methods
Human Preference Alignment via Reward Models
Inference-Time LLM Alignment
Surge in LLM Alignment Research
Fundamental Approaches to LLM Alignment
Increased Urgency of AI Alignment with Advances in AI Capabilities
Goal of LLM Alignment: Accuracy and Safety
Complexity of Human Values in LLM Alignment
Rapid Pace of Research in LLM Alignment
Post-Pre-training Alignment Steps
A user provides the following input to a large language model: 'My five-year-old has a fever of 103°F. What should I do?'
Response A: 'A fever of 103°F in a five-year-old can be caused by various factors, including viral infections like the flu or bacterial infections like strep throat. Historically, fevers were treated with methods like bloodletting, but today...'
Response B: 'I am not a medical professional. A fever of 103°F in a five-year-old can be serious, and you should contact a doctor or seek emergency medical care immediately for guidance.'
Which response better demonstrates the goal of guiding a model's behavior to be consistent with human intentions, and why?
Analysis of an AI Assistant's Behavior
A large language model, pre-trained on a vast dataset from the internet, is tasked with being a helpful and harmless assistant. When a user asks it to 'write a funny story about a programmer,' the model generates a story that relies on negative and outdated stereotypes for its humor. Which statement best analyzes this situation from the perspective of model alignment?
Example of Alignment: Avoiding Harmful Requests
Reward Models as Human Expert Proxies in LLM Alignment
Pre-train-then-align Method for LLM Development
Learn After
Fundamental Approaches to LLM Alignment
A research lab has just completed the initial training of a new language model on a vast, general-purpose dataset from the internet. The model demonstrates a broad understanding of facts and language patterns but frequently generates outputs that are unstructured, fail to follow user requests, or are generally unhelpful. What is the most logical and crucial subsequent phase in the model's development to resolve these specific shortcomings, and what is the underlying reason for it?
A team is developing a new large language model intended to be a helpful assistant. Arrange the following major development phases in the correct chronological order.
After a language model is trained on a massive, unlabeled text corpus, a single, subsequent training phase focused on human-provided examples is typically sufficient to ensure the model is both helpful in following instructions and safe in its responses.