Summary of Cell Your Model: Contrastive Explanations For Large Language Models, by Ronny Luss et al.
CELL your Model: Contrastive Explanations for Large Language Models
by Ronny Luss, Erik Miehling, Amit Dhurandhar
First submitted to arxiv on: 17 Jun 2024
Categories
- Main: Computation and Language (cs.CL)
- Secondary: Artificial Intelligence (cs.AI); 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 contrastive explanation method provides insights into why large language models (LLMs) generate specific responses to given prompts by analyzing how these models would respond if the prompt were slightly modified. The approach uses a scoring function that is meaningful to users, rather than requiring a class label. A novel budgeted algorithm intelligently creates contrasts based on this scoring function while adhering to a query budget necessary for longer contexts. The method is shown to be effective in important natural language tasks such as open-text generation and chatbot conversations. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Large language models (LLMs) are super smart computers that can understand and respond to human language. They’re like super smart robots, but instead of moving around, they process text. One problem with these LLMs is that we don’t really know why they come up with certain responses to prompts. The authors of this paper came up with a new way to figure out why an LLM might give a specific response to a prompt. They do this by looking at how the model would respond if the prompt were slightly changed. This helps us understand what makes the model decide on a particular response. This method could be really useful for things like chatbots and generating text. |
Keywords
» Artificial intelligence » Prompt » Text generation
