Summary of Making New Connections: Llms As Puzzle Generators For the New York Times’ Connections Word Game, by Tim Merino et al.
Making New Connections: LLMs as Puzzle Generators for The New York Times’ Connections Word Game
by Tim Merino, Sam Earle, Ryan Sudhakaran, Shyam Sudhakaran, Julian Togelius
First submitted to arxiv on: 15 Jul 2024
Categories
- Main: Artificial Intelligence (cs.AI)
- Secondary: Computation and Language (cs.CL)
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 paper investigates the ability of Large Language Models (LLMs) from the GPT family to generate word games for human players. Specifically, it focuses on the Connections puzzle game published daily by The New York Times, which requires semantic knowledge and abstract reasoning to solve. To generate novel puzzles, a form of metacognition is needed to model downstream reasoning. The authors propose a method using Tree of Thoughts (ToT) prompting to adapt LLMs for generating Connections puzzles. A user study is conducted to evaluate the method by comparing AI-generated puzzles with published ones. Results show that LLMs are capable puzzle creators, generating diverse sets of enjoyable, challenging, and creative puzzles as judged by human users. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The paper looks at how computers can create word games for people. One game they studied is called Connections, where you find groups of four words connected by a theme. To make new puzzles, the computer needs to think like someone trying to solve it. The researchers showed that special computers called Large Language Models (LLMs) are good at making these kinds of puzzles. They made a way for LLMs to create word games using ideas from how humans think. Then they asked people to play with AI-made puzzles and ones in the paper, and the results show that the computer-made puzzles were fun and challenging. |
Keywords
» Artificial intelligence » Gpt » Prompting
