Dataset Viewer
Auto-converted to Parquet Duplicate
language
stringclasses
3 values
question
stringlengths
14
44
expected_answer
stringlengths
1
11
en
What is 2 + 2?
4
hi
2 + 2 कितना होता है?
4
es
¿Cuánto es 2 + 2?
4
en
What is the capital of India?
New Delhi
hi
भारत की राजधानी क्या है?
नई दिल्ली
es
¿Cuál es la capital de India?
Nueva Delhi
en
What is 5 multiplied by 3?
15
hi
5 गुणा 3 कितना होता है?
15
es
¿Cuánto es 5 multiplicado por 3?
15
en
Which planet is called the Red Planet?
Mars
hi
किस ग्रह को लाल ग्रह कहा जाता है?
मंगल
es
¿Qué planeta se conoce como el planeta rojo?
Marte
en
Which is the largest planet?
Jupiter
hi
सबसे बड़ा ग्रह कौन सा है?
बृहस्पति
es
¿Cuál es el planeta más grande?
Júpiter
en
How many days are in a week?
7
hi
एक सप्ताह में कितने दिन होते हैं?
7
es
¿Cuántos días hay en una semana?
7
en
In which direction does the Sun rise?
East
hi
सूर्य किस दिशा में उगता है?
पूर्व
es
¿En qué dirección sale el Sol?
Este
en
What is the chemical formula for water?
H2O
hi
पानी का रासायनिक सूत्र क्या है?
H2O
es
¿Cuál es la fórmula química del agua?
H2O
en
What is the opposite of hot?
Cold
hi
गरम का विलोम शब्द क्या है?
ठंडा
es
¿Cuál es lo contrario de caliente?
Frío
en
How many hours are in a day?
24
hi
एक दिन में कितने घंटे होते हैं?
24
es
¿Cuántas horas hay en un día?
24

Multilingual LLM Evaluation

A small evaluation dataset for comparing language models across English, Hindi, and Spanish.

Columns

  • language: language code (en, hi, or es)
  • question: question provided to the model
  • expected_answer: reference answer used for scoring

Intended use

This dataset can be used to compare model accuracy, language adherence, and response speed across languages.

Limitations

This is a small demonstration dataset and is not a comprehensive benchmark. Results should not be used as the sole basis for selecting or deploying a model.

Source code

Evaluation code: https://github.com/Sheetal-Gupta-AI/llm-evaluation

Downloads last month
19