Summary of Defining and Evaluating Decision and Composite Risk in Language Models Applied to Natural Language Inference, by Ke Shen et al.
Defining and Evaluating Decision and Composite Risk in Language Models Applied to Natural Language Inference
by Ke Shen, Mayank Kejriwal
First submitted to arxiv on: 4 Aug 2024
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 The proposed paper addresses the risks associated with large language models (LLMs) like ChatGPT, particularly the under-confidence risk. The authors define two types of risk: decision risk and composite risk. They propose a two-level inference architecture and metrics to measure these risks in discriminative and generative LLMs. The framework is tested on four natural language commonsense reasoning datasets using RoBERTa and ChatGPT, showing its practical utility. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper explores the risks of large language models like ChatGPT. It shows that these models can be too confident or not confident enough, which can lead to wrong answers. The authors come up with a way to measure these risks and test it on four different datasets. This helps us understand how well these models work and what they’re good at. |
Keywords
» Artificial intelligence » Inference
