File size: 2,266 Bytes
01fdb44 b7c245a 01fdb44 b7c245a 01fdb44 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 |
---
language:
- en
pretty_name: Bar Exam QA
license: cc-by-sa-4.0
---
`Bar Exam QA` is a dataset for legal information retrieval and question answering. See the [original releasing paper](https://reglab.github.io/legal-rag-benchmarks/) for more information.
This repo contains the public MBE subset of questions (historical exams), which we release with gold passage annotations. In our work, we also evaluate on a private test set of questions from a BarBri multistate bar exam test preparation textbook.
The dataset was created by linking multistate bar exam multiple-choice questions with supporting gold passages from primary sources (caselaw) and secondary sources (online resources, textbooks, etc.).
The gold passages were annotated by a law student, simulating a legal research workflow.
## qa
The dataset can be loaded with the `qa` configuration. Each example consists of a question, multiple-choice answers, an answer label, and a gold passage (passage ID) label.
The dataset contains 1,195 examples.
```python
dataset = load_dataset("reglab/barexam_qa", name="qa")
```
where
- `idx`: unique example ID (string)
- `prompt`: the question prompt (context before the question, can be shared between questions)
- `question`: the question
- `choice_a`: answer choice A text
- `choice_b`: answer choice B text
- `choice_c`: answer choice C text
- `choice_d`: answer choice D text
- `answer`: A, B, C, or D
- `gold_passge`: gold passage text
- `gold_idx`: passage ID of gold passage (matches with passage ID in `passages`)
## passages
The passage pool can be loaded with `passages` configuration. Each passages consists of a passage ID, the passage text, and additional metadata if applicable (e.g, case ID for passages from Caselaw Access Project).
The passage pool contains ~900K passages.
```python
dataset = load_dataset("reglab/barexam_qa", name="passages")
```
where
- `idx`: unique passage ID (string)
- `faiss_id`: unique Faiss ID (integer)
- `case_id`: case ID (from Caselaw Access Project)
- `opinion_id`: opinion ID (from Caselaw Access Project)
- `absolute_paragraph_id`: paragraph ID in case across all opinions (from Caselaw Access Project)
- `relative_paragraph_id`: paragraph ID in opinion (from Caselaw Access Project)
- `text`: passage text |