Summary of Who’s Harry Potter? Approximate Unlearning in Llms, by Ronen Eldan and Mark Russinovich
Who’s Harry Potter? Approximate Unlearning in LLMs
by Ronen Eldan, Mark Russinovich
First submitted to arxiv on: 3 Oct 2023
Categories
- Main: Computation and Language (cs.CL)
- Secondary: Artificial Intelligence (cs.AI)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary This research proposes a novel approach to “unlearning” a portion of the training data used to develop large language models (LLMs). The issue arises because these massive models are trained on internet corpora that often contain copyrighted content, raising legal and ethical concerns for developers, users, original authors, and publishers. By unlearning specific parts of the training data without retraining the model from scratch, this technique aims to address these challenges. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Imagine you’re building a huge language model using lots of internet data. Sometimes that data belongs to someone else, which can be a problem! This paper suggests a new way to remove some of that old information, so the model doesn’t rely on it anymore. It’s like erasing memories from your brain – except with words! |
Keywords
* Artificial intelligence * Language model
