Completed
- Improvement via Perplexity Ratios
Class Central Classrooms beta
YouTube videos curated by Class Central.
Classroom Contents
Extracting Training Data from Large Language Models - Paper Explained
Automatically move to the next video in the Classroom when playback concludes
- 1 - Intro & Overview
- 2 - Personal Data Example
- 3 - Eidetic Memorization & Language Models
- 4 - Adversary's Objective & Outlier Data
- 5 - Ethical Hedging
- 6 - Two-Step Method Overview
- 7 - Perplexity Baseline
- 8 - Improvement via Perplexity Ratios
- 9 - Weights for Patterns & Weights for Memorization
- 10 - Analysis of Main Results
- 11 - Mitigation Strategies
- 12 - Conclusion & Comments