michaelkrumdickkensho
commited on
Commit
•
0a793f2
1
Parent(s):
19028d8
Update README.md
Browse files
README.md
CHANGED
@@ -36,21 +36,33 @@ dataset_info:
|
|
36 |
|
37 |
|
38 |
<p align="left">
|
39 |
-
<img src="bizbench_pyramid.png"
|
40 |
</p>
|
41 |
|
42 |
# BizBench: A Quantitative Reasoning Benchmark for Business and Finance
|
43 |
|
44 |
Public dataset for [BizBench](https://arxiv.org/abs/2311.06602).
|
45 |
|
46 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
47 |
|
48 |
| Dataset | Train/Few Shot Data | Test Data |
|
49 |
| --- | --- | --- |
|
50 |
| **Program Synthesis** | | |
|
51 |
| FinCode | 7 | 47 |
|
52 |
| CodeFinQA | 4668 | 795 |
|
53 |
-
| CodeTATQA |
|
54 |
| **Quantity Extraction** | | |
|
55 |
| ConvFinQA (E) | | 629 |
|
56 |
| TAT-QA (E) | | 120 |
|
|
|
36 |
|
37 |
|
38 |
<p align="left">
|
39 |
+
<img src="bizbench_pyramid.png">
|
40 |
</p>
|
41 |
|
42 |
# BizBench: A Quantitative Reasoning Benchmark for Business and Finance
|
43 |
|
44 |
Public dataset for [BizBench](https://arxiv.org/abs/2311.06602).
|
45 |
|
46 |
+
Answering questions within business and finance requires reasoning, precision, and a wide-breadth of technical knowledge.
|
47 |
+
Together, these requirements make this domain difficult for large language models (LLMs).
|
48 |
+
We introduce BizBench, a benchmark for evaluating models' ability to reason about realistic financial problems.
|
49 |
+
BizBench comprises **eight quantitative reasoning tasks**, focusing on question-answering (QA) over financial data via program synthesis.
|
50 |
+
We include three financially-themed code-generation tasks from newly collected and augmented QA data.
|
51 |
+
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.
|
52 |
+
Collectively, these tasks evaluate a model's financial background knowledge, ability to parse financial documents, and capacity to solve problems with code.
|
53 |
+
We conducted an in-depth evaluation of open-source and commercial LLMs, comparing and contrasting the behavior of code-focused and language-focused models.
|
54 |
+
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.
|
55 |
+
|
56 |
+
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).
|
57 |
+
|
58 |
+
## Dataset Statistics
|
59 |
|
60 |
| Dataset | Train/Few Shot Data | Test Data |
|
61 |
| --- | --- | --- |
|
62 |
| **Program Synthesis** | | |
|
63 |
| FinCode | 7 | 47 |
|
64 |
| CodeFinQA | 4668 | 795 |
|
65 |
+
| CodeTATQA | 2856 | 2000 |
|
66 |
| **Quantity Extraction** | | |
|
67 |
| ConvFinQA (E) | | 629 |
|
68 |
| TAT-QA (E) | | 120 |
|