Learn Before
Preference Notation in Human Feedback
In the context of human feedback for language models, the notation is used to formally represent a preference. It signifies that a human annotator has judged output to be of higher quality or more desirable than output .
0
1
Tags
Ch.4 Alignment - Foundations of Large Language Models
Foundations of Large Language Models
Foundations of Large Language Models Course
Computing Sciences
Related
Evaluation Criteria for Pairwise Comparison in RLHF
Bradley-Terry Model
Reward Model Training as a Ranking Problem in RLHF
Listwise Ranking for Human Feedback in RLHF
Importance of Variability in Pairwise Preference Data
Evaluating a Feedback Collection Strategy
A development team is refining a language model's ability to generate summaries. For each source document, they have the model produce two different summaries. They then present these two summaries side-by-side to a human annotator and ask them to select the one that is of higher quality. Which statement best analyzes the primary strength of this specific approach for collecting human feedback?
Rationale for a Feedback Collection Method
Binary Encoding of Pairwise Feedback in RLHF
Preference Notation in Human Feedback
Learn After
Example of a Human Preference Ranking in RLHF
Ranked Preference Notation
Example of Listwise Ranking in RLHF
A language model generates two different summaries for a given article: Summary 1 and Summary 2. A human evaluator is tasked with reviewing them and determines that Summary 1 is more coherent and factually accurate than Summary 2. How would this specific judgment be formally expressed using standard preference notation?
A human annotator provides the following judgments for four text completions (C1, C2, C3, C4) generated in response to a single prompt: C1 ≻ C4, C4 ≻ C2, and C2 ≻ C3. Based on this information, arrange the completions in order from most preferred to least preferred.
Limitations of Preference Notation