Training Data for a Sentence-Level Fact-Checker
Describe how the human annotators should structure each labeled example in the training dataset. Specifically, what two components should each training instance consist of to enable a model to learn to assign a direct, numerical 'inaccuracy score' to any given sentence?
0
1
Tags
Ch.4 Alignment - Foundations of Large Language Models
Foundations of Large Language Models
Foundations of Large Language Models Course
Computing Sciences
Application in Bloom's Taxonomy
Cognitive Psychology
Psychology
Social Science
Empirical Science
Science
Related
Automated Segment Scoring via LLM-Generated Ratings
A development team is building a system to automatically flag individual user comments for toxicity. They have a large dataset where each comment has been rated by a human moderator on a scale of 1 (not toxic) to 5 (highly toxic). Which of the following is the most direct and suitable method for training a model to assign a toxicity rating to each new comment?
Training Data for a Sentence-Level Fact-Checker
Justifying a Modeling Approach for Fact-Checking