Improve knights-and-knaves dataset card

README.md

@@ -1,9 +1,12 @@
 ---
+language:
+- en
 license: cc-by-nc-sa-4.0
+size_categories:
+- 1K<n<10K
 task_categories:
 - question-answering
-language:
-- en
+pretty_name: Knights and Knaves Logical Reasoning Benchmark
 configs:
 - config_name: train
   data_files:
@@ -54,16 +57,14 @@ configs:
 tags:
 - logical
 - reasoning
-pretty_name: K
-size_categories:
-- 1K<n<10K
+- knights-and-knaves
+- memorization
 ---
 
+# Knights and Knaves Logical Reasoning Benchmark [[Project Page]](https://memkklogic.github.io/)
 
+This dataset provides a dynamically generated benchmark for evaluating the logical reasoning capabilities of Large Language Models (LLMs), with a specific focus on quantifying memorization effects.  The benchmark is based on Knights and Knaves puzzles of varying complexity, allowing for a nuanced investigation of how LLMs balance memorization and genuine reasoning.  A key feature is the ability to assess generalization by introducing perturbations to the training puzzles.
 
-# 📘 knights-and-knaves Dataset [[Project Page]](https://memkklogic.github.io/)
-
-The **knights-and-knaves dataset** serves as a logical reasoning benchmark to evaluate the reasoning capabilities of LLMs.
 
 **🚀🚀 Check out the [perturbed knights-and-knaves dataset](https://huggingface.co/datasets/K-and-K/perturbed-knights-and-knaves) to evaluate the memorization of LLMs in reasoning.** 
 
@@ -75,16 +76,17 @@ To load the dataset:
 from datasets import load_dataset
 data_subject = load_dataset('K-and-K/knights-and-knaves','test',split="2ppl")
 ```
-* Available subset: `test`, `train`.
-* Available split: `2ppl`,`3ppl`,`4ppl`,`5ppl`,`6ppl`,`7ppl`,`8ppl`.
+* Available subsets: `test`, `train`.
+* Available splits: `2ppl`,`3ppl`,`4ppl`,`5ppl`,`6ppl`,`7ppl`,`8ppl`.  (Number of people involved in the puzzle)
 
-## 🛠️ Codebase
+## 🛠️ Codebase and Paper
 
-To evaluate LLMs on our datasets, visit our [GitHub repository](https://github.com/AlphaPav/mem-kk-logic/).
+For detailed evaluation methodologies, fine-tuning procedures, and in-depth analysis, refer to our [GitHub repository](https://github.com/AlphaPav/mem-kk-logic/) and the accompanying paper:
+
+[On Memorization of Large Language Models in Logical Reasoning](https://hf.co/papers/2410.23123)
 
-## ⭐ Citing our Work
 
-If you find our codebase and datasets beneficial, kindly cite our work:
+## ⭐ Citing our Work
 
 ```bibtex
 @article{xie2024memorization,