kensho/bizbench
收藏资源简介:
--- license: apache-2.0 configs: - config_name: default data_files: - split: train path: data/train-* - split: test path: data/test-* dataset_info: features: - name: question dtype: string - name: answer dtype: string - name: task dtype: string - name: context dtype: string - name: context_type dtype: string - name: options sequence: string - name: program dtype: string splits: - name: train num_bytes: 52823429 num_examples: 14377 - name: test num_bytes: 15720371 num_examples: 4673 download_size: 23760863 dataset_size: 68543800 --- <p align="left"> <img src="bizbench_pyramid.png"> </p> # BizBench: A Quantitative Reasoning Benchmark for Business and Finance Public dataset for [BizBench](https://arxiv.org/abs/2311.06602). Answering questions within business and finance requires reasoning, precision, and a wide-breadth of technical knowledge. Together, these requirements make this domain difficult for large language models (LLMs). We introduce BizBench, a benchmark for evaluating models' ability to reason about realistic financial problems. BizBench comprises **eight quantitative reasoning tasks**, focusing on question-answering (QA) over financial data via program synthesis. We include three financially-themed code-generation tasks from newly collected and augmented QA data. Additionally, we isolate the reasoning capabilities required for financial QA: reading comprehension of financial text and tables for extracting intermediate values, and understanding financial concepts and formulas needed to calculate complex solutions. Collectively, these tasks evaluate a model's financial background knowledge, ability to parse financial documents, and capacity to solve problems with code. We conducted an in-depth evaluation of open-source and commercial LLMs, comparing and contrasting the behavior of code-focused and language-focused models. We demonstrate that the current bottleneck in performance is due to LLMs' limited business and financial understanding, highlighting the value of a challenging benchmark for quantitative reasoning within this domain. We have also develop a heavily curated leaderboard with a held-out test set open to submission: [https://benchmarks.kensho.com/](https://benchmarks.kensho.com/). This set was manually curated by financial professionals and further cleaned by hand in order to ensure the highest quality. A sample pipeline for using this dataset can be found at [https://github.com/kensho-technologies/benchmarks-pipeline](https://github.com/kensho-technologies/benchmarks-pipeline). ## Dataset Statistics | Dataset | Train/Few Shot Data | Test Data | | --- | --- | --- | | **Program Synthesis** | | | | FinCode | 7 | 47 | | CodeFinQA | 4668 | 795 | | CodeTATQA | 2856 | 2000 | | **Quantity Extraction** | | | | ConvFinQA (E) | | 629 | | TAT-QA (E) | | 120 | | SEC-Num | 6846 | 2000 | | **Domain Knowledge** | | | | FinKnow | | 744 | | ForumlaEval | | 50 |
license: Apache-2.0开源许可证 configs: - config_name: 默认配置 data_files: - split(拆分集): 训练集(train) path: data/train-* - split(拆分集): 测试集(test) path: data/test-* dataset_info(数据集信息): features(特征): - name: 问题(question) dtype(数据类型): 字符串(string) - name: 答案(answer) dtype(数据类型): 字符串(string) - name: 任务类型(task) dtype(数据类型): 字符串(string) - name: 上下文(context) dtype(数据类型): 字符串(string) - name: 上下文类型(context_type) dtype(数据类型): 字符串(string) - name: 选项(options) sequence(序列类型): 字符串(string) - name: 程序(program) dtype(数据类型): 字符串(string) splits(拆分集): - name: 训练集(train) num_bytes: 52823429 num_examples: 14377 - name: 测试集(test) num_bytes: 15720371 num_examples: 4673 download_size: 23760863 dataset_size: 68543800 <p align="left"> <img src="bizbench_pyramid.png"> </p> # BizBench:面向商务与金融领域的定量推理基准数据集 本数据集为[BizBench](https://arxiv.org/abs/2311.06602)的公开数据集。 商务与金融领域的问题解答需要推理能力、严谨性以及广博的专业技术知识。这些综合要求使得该领域成为大语言模型(Large Language Model, LLM)的难点场景。为此我们推出BizBench基准数据集,用于评估模型针对真实金融问题的推理能力。 BizBench包含**8项定量推理任务**,核心为通过程序合成(program synthesis)完成金融数据相关的问答(Question-Answering, QA)任务。我们从新采集并增强的问答数据中,新增了3项金融主题的代码生成任务。 此外,我们将金融问答所需的推理能力进行拆解:包括对金融文本与表格的阅读理解以提取中间数值,以及理解用于求解复杂问题的金融概念与公式。综合来看,这些任务可评估模型的金融背景知识储备、金融文档解析能力,以及通过代码求解问题的能力。 我们针对开源与商用大语言模型开展了深度评估,对比分析了专注代码与专注自然语言的模型的表现差异。我们的研究表明,当前模型性能的瓶颈源于大语言模型对商务与金融领域的理解能力不足,这也凸显了该领域定量推理基准数据集的重要价值。 我们还搭建了经过严格筛选的排行榜,其预留测试集对外开放提交:[https://benchmarks.kensho.com/](https://benchmarks.kensho.com/)。该测试集由金融专业人士手动筛选,并经人工进一步清洗,以确保最高的数据质量。本数据集的使用示例流程可参考:[https://github.com/kensho-technologies/benchmarks-pipeline](https://github.com/kensho-technologies/benchmarks-pipeline)。 ## 数据集统计信息 | 数据集 | 训练/少样本数据 | 测试数据 | | --- | --- | --- | | **程序合成(program synthesis)** | | | | FinCode | 7 | 47 | | CodeFinQA | 4668 | 795 | | CodeTATQA | 2856 | 2000 | | **数值提取(quantity extraction)** | | | | ConvFinQA (E) | | 629 | | TAT-QA (E) | | 120 | | SEC-Num | 6846 | 2000 | | **领域知识(domain knowledge)** | | | | FinKnow | | 744 | | ForumlaEval | | 50 |
BizBench 数据集概述
数据集信息
许可证
- Apache 2.0
配置
- 默认配置
- 训练数据路径:
data/train-* - 测试数据路径:
data/test-*
- 训练数据路径:
特征
question: 字符串answer: 字符串task: 字符串context: 字符串context_type: 字符串options: 字符串序列program: 字符串
数据分割
- 训练集
- 字节数: 52823429
- 样本数: 14377
- 测试集
- 字节数: 15720371
- 样本数: 4673
数据大小
- 下载大小: 23760863 字节
- 数据集大小: 68543800 字节
数据集统计
| 数据集 | 训练/小样本数据 | 测试数据 |
|---|---|---|
| 程序合成 | ||
| FinCode | 7 | 47 |
| CodeFinQA | 4668 | 795 |
| CodeTATQA | 2856 | 2000 |
| 数量提取 | ||
| ConvFinQA (E) | 629 | |
| TAT-QA (E) | 120 | |
| SEC-Num | 6846 | 2000 |
| 领域知识 | ||
| FinKnow | 744 | |
| ForumlaEval | 50 |




