sileod/mindgames

---
language:
- en
license: apache-2.0
multilinguality:
- monolingual
task_categories:
- text-classification
task_ids:
- natural-language-inference
- multi-input-text-classification
tags:
- theory of mind
- tom
- Logical-Reasoning
- Modal-Logic
- Reasoning
- Logics
- Logic
- nli
- model-checking
- natural language inference
dataset_info:
features:
- name: premise
dtype: string
- name: smcdel_problem
dtype: string
- name: n_announcements
dtype: int64
- name: pbcheck
dtype: string
- name: hypothesis
dtype: string
- name: setup
dtype: string
- name: hypothesis_depth
dtype: int64
- name: n_agents
dtype: int64
- name: label
dtype: string
- name: names
sequence: string
- name: index
dtype: int64
- name: s-l
dtype: string
- name: deberta_pred
dtype: int64
- name: deberta_confidence
dtype: float64
- name: difficulty
dtype: float64
splits:
- name: train
num_bytes: 8702021
num_examples: 11174
- name: validation
num_bytes: 2904084
num_examples: 3725
- name: test
num_bytes: 2909341
num_examples: 3725
download_size: 2989857
dataset_size: 14515446
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*
- split: test
path: data/test-*
---
Mindgame dataset

Code:
https://github.com/sileod/llm-theory-of-mind

Article (Accepted at EMNLP 2023 Findings):
https://arxiv.org/abs/2305.03353
```
@article{sileo2023mindgames,
title={MindGames: Targeting Theory of Mind in Large Language Models with Dynamic Epistemic Modal Logic},
author={Sileo, Damien and Lernould, Antoine},
journal={arXiv preprint arXiv:2305.03353},
year={2023}
}
```

原始信息汇总

数据集概述

基本信息

• 语言: 英语
• 许可证: Apache-2.0
• 多语言性: 单语
• 任务类别: 文本分类
• 任务ID:
  • 自然语言推理
  • 多输入文本分类
• 标签:
  • 理论思维
  • 逻辑推理
  • 模态逻辑
  • 推理
  • 逻辑
  • 自然语言推理
  • 模型检查

数据集特征

• premise: 字符串
• smcdel_problem: 字符串
• n_announcements: 整数
• pbcheck: 字符串
• hypothesis: 字符串
• setup: 字符串
• hypothesis_depth: 整数
• n_agents: 整数
• label: 字符串
• names: 字符串序列
• index: 整数
• s-l: 字符串
• deberta_pred: 整数
• deberta_confidence: 浮点数
• difficulty: 浮点数

数据集分割

• 训练集:
  • 字节数: 8702021
  • 示例数: 11174
• 验证集:
  • 字节数: 2904084
  • 示例数: 3725
• 测试集:
  • 字节数: 2909341
  • 示例数: 3725

数据集大小

• 下载大小: 2989857
• 数据集大小: 14515446

配置

• 默认配置:
  • 训练数据路径: data/train-*
  • 验证数据路径: data/validation-*
  • 测试数据路径: data/test-*