Summary of Identifying the Source Of Generation For Large Language Models, by Bumjin Park and Jaesik Choi
Identifying the Source of Generation for Large Language Models
by Bumjin Park, Jaesik Choi
First submitted to arxiv on: 5 Jul 2024
Categories
- Main: Computation and Language (cs.CL)
- Secondary: Machine Learning (cs.LG)
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 The proposed solution introduces token-level source identification in the decoding step to map token representations to reference documents, enabling users to obtain hints about reliability and potential privacy infringements. The bi-gram source identifier, a multi-layer perceptron with two successive token representations as input, is designed for better generalization. Experiments on Wikipedia and PG19 datasets with various Large Language Models (LLMs), layer locations, and identifier sizes demonstrate the feasibility of this approach. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Large language models memorize text from several sources, but they don’t remember where the text comes from. This makes it hard to know if what they generate is true or not. The problem is even bigger when it comes to privacy because you can’t trust what they say. To fix this, researchers came up with an idea to add a special code to each word that says which document it’s from. They tested this on two big datasets and found that it works. |
Keywords
» Artificial intelligence » Generalization » Token
