---
license: apache-2.0
language:
- my
pretty_name: Myanmar G12L Benchmar
size_categories:
- n<1K
dataset_info:
  features:
  - name: title
    dtype: string
  - name: question
    dtype: string
  - name: answer
    dtype: string
  - name: type
    dtype: string
  - name: option_a
    dtype: string
  - name: option_b
    dtype: string
  - name: option_c
    dtype: string
  splits:
  - name: test
    num_bytes: 1643102
    num_examples: 962
  download_size: 469838
  dataset_size: 1643102
configs:
- config_name: default
  data_files:
  - split: test
    path: data/test-*
---

# Myanmar G12L Benchmark

<!-- Provide a quick summary of the dataset. -->

Burmese language matriculation examination questions to benchmark formal knowledge in literature.

## Dataset Details

Our Burmese language matriculation examination resource is a comprehensive tool for evaluating and strengthening formal literary knowledge. 
It features the following question formats: Short Answer, True or False, Metaphor Analysis, Fill-in-the-Blank, Multiple Choice, Long-Form Response, and Meaning Interpretation. 
Each question has been extracted from past examination papers and authoritative exam guides

- **Curated by:** [Pyae Sone Myo](https://www.linkedin.com/pyaesonemyo), [Min Thein Kyaw](https://www.instagram.com/jerrytheinkyaw?igsh=OGQ5ZDc2ODk2ZA==), [May Myat Noe Aung](https://www.linkedin.com/in/may-myat-noe-aung/), [Arkar Zaw](https://linkedin.com/in/ar-kar-zaw-6885b720b)
- **Language(s) (NLP):** Burmese
- **License:** Apache 2.0


## Evaluation

<!-- Provide the basic links for the dataset. -->
To evaluate models on this benchmark, you can use the `ayamytk` (Aya Myanmar Toolkit) that was originally developed for running this benchmark.

1. Install the toolkit directly.
```py
!pip install git+https://github.com/Rickaym/aya-my-tk
```

2. Run the `ExamEval`
```py
from ayamytk.test.bench import evals
from ayamytk.test.bench.sampler.custom_sampler import CustomSampler

def chat(messages):
  # Add your inference code here
  return ...

evals.run(samplers={"your-model": CustomSampler(chat=chat)}, evals="mg12l")
```


## Dataset Structure

**Schema overview**
This dataset captures individual exam items with seven core fields:

* **title**: a brief non-unique identifier for the question
* **question**: the full prompt or stem
* **answer**: the correct response
* **type**: one of `MCQ`, `TOF`, `FIB`, `SHORT_QNA`, `LONG_QNA`, `MEANING_QNA`, or `METAPHOR_QNA`
* **option\_a**, **option\_b**, **option\_c**: the three distractors for multiple-choice items (populated only when `type = MCQ`; otherwise left blank)

Each row corresponds to a single question, and non-MCQ entries simply omit the option fields.

### Source Data

#### Data Collection and Processing

Steps:
1. Google Document OCR
2. Manual Extraction and Correction


## Bias, Risks, and Limitations

<!-- This section is meant to convey both technical and sociotechnical limitations. -->

This dataset only captures the literature subject of the matriculation exam.