id stringlengths 11 64 | domain stringclasses 6
values | subject stringclasses 11
values | category stringlengths 7 46 | grade stringclasses 7
values | answer_type stringclasses 2
values | passage stringlengths 0 4.13k | question stringlengths 31 978 | options listlengths 0 5 | answer_index int32 -1 4 | answer stringlengths 1 99 | explanation stringlengths 27 9.12k | images listlengths 0 6 | origin_source stringclasses 1
value |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
education_research_and_advanced_science_001 | Education, Research & Advanced Science | Engineering | A.C Fundamentals, Circuits And Circuit Theory | Graduate | single_select | The purpose of a parallel circuit resonance is to magnify | [
"Current",
"Voltage",
"Power",
"Frequency"
] | 1 | Voltage | In a parallel circuit resonance, the purpose is to <strong>magnify voltage</strong>.
<br/><br/>Resonance in a parallel circuit occurs when the impedance of the circuit becomes purely resistive, causing the current to be magnified and the voltage across the circuit elements to be maximized.
<br/><br/><strong>Option A:</... | [] | EduGorilla | |
education_research_and_advanced_science_002 | Education, Research & Advanced Science | Engineering | Analog Electronics | Graduate | single_select | In a CB amplifier the maximum efficiency could be | [
"99%",
"85%",
"50%",
"25%"
] | 2 | 50% | <strong>Option A: 99%</strong><br/>
An efficiency of 99% is highly unlikely for a Common Base (CB) amplifier. Such a high efficiency is typically not achievable in practical amplifier circuits due to inherent losses in the components and design. Therefore, this option is <strong>incorrect</strong>.<br/><br/>
<strong>Op... | [] | EduGorilla | |
education_research_and_advanced_science_003 | Education, Research & Advanced Science | Engineering | Artificial Intelligence | Graduate | single_select | To invoke the LISP system, you must enter | [
"AI",
"LISP",
"CL (Common Lisp)",
"both b and c",
"None of the above"
] | 3 | both b and c | To invoke the LISP system, you can use either the keyword <strong>AI</strong> or <strong>LISP</strong>, or you can enter <strong>CL (Common Lisp)</strong>. These are different ways to initiate the LISP environment.<br/>
Option A: <strong>AI</strong> is one of the keywords used to invoke the LISP system.<br/>
Option B: ... | [] | EduGorilla | |
education_research_and_advanced_science_004 | Education, Research & Advanced Science | Engineering | Artificial Intelligence | Graduate | single_select | DEC advertises that it helped to create "the world's first expert system routinely used in an industrial environment," called XCON or: | [
"PDP-11",
"Rl",
"VAX",
"MAGNOM",
"None of the above"
] | 2 | VAX | DEC (Digital Equipment Corporation) advertised that it aided in the development of "the world's first expert system routinely used in an industrial environment," known as XCON. XCON was developed by researchers at Carnegie Mellon University in the 1970s to configure orders for DEC's VAX computers.<br/>
Option A: PDP-11... | [] | EduGorilla | |
education_research_and_advanced_science_005 | Education, Research & Advanced Science | Engineering | Automobile Engineering | Graduate | single_select | The antiknock property of compression ignition engine fuel can be improved by adding | [
"Tetraethyl lead",
"Trimethyl pentane",
"Amyl nitrate",
"Hexadecane"
] | 2 | Amyl nitrate | The antiknock property of compression ignition engine fuel can be improved by adding Amyl nitrate | [] | EduGorilla | |
education_research_and_advanced_science_006 | Education, Research & Advanced Science | Engineering | Automobile Engineering | Graduate | single_select | A basic difference between the spark ignition engine and the Diesel engine is that the | [
"Diesel engine compresses air alone instead of an air-fuel mixture",
"Air temperature ignites the fuel in the Diesel engine",
"Fuel is sprayed into the combustion chamber in the Diesel engine as the piston nears T.D.C. on the compression stroke",
"All of the above"
] | 3 | All of the above | A basic difference between the spark ignition engine and the Diesel engine is that the:<br/>
- Diesel engine compresses air alone instead of an air-fuel mixture<br/>
- Air temperature ignites the fuel in the Diesel engine<br/>
- Fuel is sprayed into the combustion chamber in the Diesel engine as the piston nears T.D.C.... | [] | EduGorilla | |
education_research_and_advanced_science_007 | Education, Research & Advanced Science | Biology | Biology | Others | single_select | A viral species is a group of viruses that has | [
"Same ecological niche",
"Same genetic information",
"Both A & B",
"None of the above"
] | 2 | Both A & B | A viral species is a group of viruses that has same ecological niche and same genetic information. | [] | EduGorilla | |
education_research_and_advanced_science_008 | Education, Research & Advanced Science | Biology | Biology | Others | single_select | Fossils of blue green algae were found in lime stones with the age almost as old as | [
"1.3 Billion years",
"2.3 Billion years",
"3.3 Billion years",
"4.3 Billion years"
] | 1 | 2.3 Billion years | Fossils of blue green algae were found in lime stones with the age almost as old as 2.3 Billion years. | [] | EduGorilla | |
education_research_and_advanced_science_009 | Education, Research & Advanced Science | Biology | Biology | Others | single_select | Microscopic organism which passively float on the surface of water is called | [
"Phytoplanktons",
"Zooplanktons",
"Both (a) and (b) both",
"None to these"
] | 2 | Both (a) and (b) both | Microscopic organism which passively float on the surface of water is called Zooplanktons and Phytoplanktons. | [] | EduGorilla | |
education_research_and_advanced_science_010 | Education, Research & Advanced Science | Biology | Biology | Others | single_select | These are rods or filament like very small bodies present in the cytoplasm of all cells of plants and animals. | [
"Golgi Bodies",
"Ribosomes",
"Lysosomes",
"Mitochondria"
] | 3 | Mitochondria | Mitochondria are rods or filament like very small bodies present in the cytoplasm of all cells of plants and animals. | [] | EduGorilla | |
enterprise_operations_professional_services_and_supply_chain_001 | Enterprise Operations, Professional Services & Supply Chain | Commerce | Quantitative Aptitude | Senior High school / Grade 12 | single_select | <p>Look at the pi<span>cture given below-</span></p><img alt="" src="image_0"/><span><br/></span><p>This graph shows that:</p><p>1. There is a shift in the demand curve</p><p>2. Price is assumed to be constant at this moment</p><p>3. There is an upward shift in the demand curve</p><p>4. Other factors are assumed to be ... | [
"(3) and (4)",
"(1), (2), and (5)",
"(3), (4), and (5)",
"(1) and (4)"
] | 1 | (1), (2), and (5) | <p>(1), (2), and (5)</p><p>1. There is a shift in the demand curve</p><p>2. Price is assumed to be constant at this moment</p><p>5. An increase in the demand</p> | [
{
"bytes": [
137,
80,
78,
71,
13,
10,
26,
10,
0,
0,
0,
13,
73,
72,
68,
82,
0,
0,
1,
127,
0,
0,
1,
52,
8,
6,
0,
0,
0,
196,
219,... | EduGorilla | |
enterprise_operations_professional_services_and_supply_chain_002 | Enterprise Operations, Professional Services & Supply Chain | Commerce | Quantitative Aptitude | Senior High school / Grade 12 | single_select | <p><span>Refer to the graph given below-</span><span></span></p><img alt="" src="image_0"/><span><br/></span><p>The above graph represents that-</p><p>1. Assume that the quantity of money remains unchanged.</p><p>2. The rise in the money demand or liquidity preference curve from LP1 to LP2.</p><p>3. It is worth noting ... | [
"(1), (2), (3), and (4)",
"(1), (2), and (4)",
"(3) and (4)",
"All of the above"
] | 1 | (1), (2), and (4) | Point 3. It is worth noting that when the liquidity preference curve rises from LP1 to LP2, the amount of money held does not increase (not decrease). | [
{
"bytes": [
137,
80,
78,
71,
13,
10,
26,
10,
0,
0,
0,
13,
73,
72,
68,
82,
0,
0,
1,
130,
0,
0,
1,
72,
8,
0,
0,
0,
0,
120,
226,... | EduGorilla | |
enterprise_operations_professional_services_and_supply_chain_003 | Enterprise Operations, Professional Services & Supply Chain | Commerce | Numerical Ability | Graduate | single_select | From following table find number of accounts that have suffered Risk rating migration during 2017-18 <br/><span><span><span><img src="image_0" style="display: inline"/></span></span></span> | [
"2",
"19",
"21",
"25",
"None of the above"
] | 1 | 19 | <span><span style="line-height: 150%">Credit rating model essentially differentiates borrowers, based on the degree of stability in terms of top line and bottom line revenue generation. This is because where uncertainty in revenue generation in a business is more chances of failing in keeping financial commitments to t... | [
{
"bytes": [
137,
80,
78,
71,
13,
10,
26,
10,
0,
0,
0,
13,
73,
72,
68,
82,
0,
0,
2,
134,
0,
0,
0,
87,
8,
0,
0,
0,
0,
188,
28,
... | EduGorilla | |
enterprise_operations_professional_services_and_supply_chain_004 | Enterprise Operations, Professional Services & Supply Chain | Management | Numerical Ability | Graduate | single_select | <p>Expected Cash flow from a project are as follows: </p><img src="image_0"/> | If the cost of capital is 12%, then from the cash flows mentioned at the above question, calculate NPV. | [
"Rs.1300",
"Rs.8040",
"Rs.9040",
"Rs.10000",
"Rs. 11400"
] | 2 | Rs.9040 | <p><span><span style="line-height: 150%">(initial cash flow)+ {Periodic cash flow /(1+% cost of capital)<sup>n</sup>, Periodic cash flow /(1+% cost of capital)<sup>n+1</sup>, Periodic cash flow /(1+% cost of capital)<sup>n+2</sup>…………..}</span></span></p> <span><span style="line-height: 150%"><span><img src="image_1" s... | [
{
"bytes": [
137,
80,
78,
71,
13,
10,
26,
10,
0,
0,
0,
13,
73,
72,
68,
82,
0,
0,
1,
46,
0,
0,
0,
231,
8,
2,
0,
0,
0,
47,
178,
... | EduGorilla |
enterprise_operations_professional_services_and_supply_chain_005 | Enterprise Operations, Professional Services & Supply Chain | Management | Numerical Ability | Graduate | single_select | <p>Expected Cash flow from a project are as follows: </p><img src="image_0"/> | Working capital leverage depends on ________. | [
"% change in Return on Investment",
"% change in Return on Inventory",
"% Change in Current assets",
"Both A and C",
"Both A and B"
] | 3 | Both A and C | <span><span style="line-height: 150%">Working capital leverage= % Change in Return on Investment/% change in Current Assets</span></span><span><span style='font-size:16px;line-height:150%;font-family: "Verdana",sans-serif;color:black'></span></span> | [
{
"bytes": [
137,
80,
78,
71,
13,
10,
26,
10,
0,
0,
0,
13,
73,
72,
68,
82,
0,
0,
1,
46,
0,
0,
0,
231,
8,
2,
0,
0,
0,
47,
178,
... | EduGorilla |
enterprise_operations_professional_services_and_supply_chain_006 | Enterprise Operations, Professional Services & Supply Chain | Management | English Language | Graduate | single_select | <p><b><span>Direction</span>: </b>Read the given information and answer the questions that follow. </p><p>The table shown below represents cash flows of an X% annual coupon G-sec having face value of INR 1000. </p><img src="image_1"/> | <span> <span><b><span>Direction</span></b><span>: Read the given information and answer the questions that follow.</span></span> <br/><span><span><span><img src="image_0" style="display: inline"/></span></span></span> <br/><span><span>It is also given that revenue from operations for the same year was INR 3,200,000</sp... | [
"2.29:1",
"2.33:1",
"3:2.5",
"1.5:1.1",
"None of these"
] | 0 | 2.29:1 | <p><span><span style="line-height: 150%">Capital Employed = Reserves and Surplus + Share Capital + Long-term Debts = (100,000 + 440,000) + (200,000 + 300,000) + (250,000 + 130,000) = 1,400,000</span></span></p> <p><span><span style="line-height: 150%">Capital Employed Turnover Ratio = Revenue from operations/Capital Em... | [
{
"bytes": [
137,
80,
78,
71,
13,
10,
26,
10,
0,
0,
0,
13,
73,
72,
68,
82,
0,
0,
2,
154,
0,
0,
0,
221,
8,
2,
0,
0,
0,
148,
172... | EduGorilla |
enterprise_operations_professional_services_and_supply_chain_007 | Enterprise Operations, Professional Services & Supply Chain | Management | General/Economy/Banking Awareness | Graduate | single_select | <p><span>Direction</span>: Read the given information and answer the questions that follow. </p><img src="image_0"/> <p>From above give data:<br/></p> | What is the quick ratio of the firm? | [
"0.83:1",
"0.89:1",
"0.78:1",
"0.75:1",
"None of these"
] | 2 | 0.78:1 | <p><span>Quick ratio = Quick assets / Current liabilities</span></p> <p><span>Quick assets = current assets – (Inventories + advanced tax) = 140, 000 – (60,000 + 5000)</span></p> <p><span>Quick assets = INR 75, 000</span></p> <p><span>Current liabilities = Trade payables + short term borrowings</span></p> <p><span>= 90... | [
{
"bytes": [
137,
80,
78,
71,
13,
10,
26,
10,
0,
0,
0,
13,
73,
72,
68,
82,
0,
0,
1,
24,
0,
0,
0,
129,
8,
0,
0,
0,
0,
62,
251,
... | EduGorilla |
enterprise_operations_professional_services_and_supply_chain_008 | Enterprise Operations, Professional Services & Supply Chain | Management | English | Graduate | single_select | <p><b>Direction:</b> Read the given table carefully and answer the following questions.</p><p>The table given below shows the sum of money borrowed by Ajay, Amar, Arun, Aman, and Anil on a certain rate for a certain period of time either on simple interest or compound interest. Some data are missing.</p><p><img alt="" ... | The ratio of Simple interest paid by Arun and Anil is 3 ∶ 4. Find the rate of interest paid by Arun if the time period for both Arun and Anil is the same? | [
"11.50%",
"10.75%",
"11.25%",
"9.75%"
] | 2 | 11.25% | <p>The principal amount borrowed by Arun is Rs. 12000 and he paid Rs. 900 as a simple interest</p><p>And Principal amount borrowed by Anil is Rs. 18000 at a 10% rate of interest.</p><p>Ratio of Simple interest paid by Arun to Anil is 3 ∶ 4</p><p>Let the ratio be x.</p><p>⇒ 3x = 900</p><p>⇒ x = 300</p><p>Interest paid b... | [
{
"bytes": [
137,
80,
78,
71,
13,
10,
26,
10,
0,
0,
0,
13,
73,
72,
68,
82,
0,
0,
1,
117,
0,
0,
1,
87,
8,
6,
0,
0,
0,
93,
153,
... | EduGorilla |
Multi-Domain Question Answering Dataset
A sample of multi-domain, exam-style question answering data with worked explanations. Questions span finance, quantitative aptitude, software engineering, medical and life sciences, enterprise operations, and advanced science and engineering. Many records include embedded images (charts, tables, diagrams, code screenshots) alongside the text.
This repository contains a 70-row sample. It is intended for evaluating the format, schema and quality of the data, and for small-scale experiments. A much larger corpus is available from the provider (see Contact).
Dataset Summary
| Rows | 70 |
| Columns | 14 |
| Language | English |
| Format | Parquet (single train split) |
| Question types | Single-select multiple choice (68), free-form numeric/text input (2) |
| Images | 74 PNG images embedded across 60 rows (about 2 MB total) |
| Source field | origin_source = EduGorilla for all rows |
Domain breakdown
| Domain | Rows | Subjects |
|---|---|---|
| Finance | 20 | Economics, Accountancy |
| Education, Research & Advanced Science | 10 | Engineering, Biology |
| Enterprise Operations, Professional Services & Supply Chain | 10 | Commerce, Management |
| Medical & Life Sciences | 10 | Nursing, Environmental Science, Medical |
| Quantitative (Quant) | 10 | Quant |
| Software Engineering | 10 | IT & Computer |
Difficulty / level (grade)
Graduate (36), Post Graduate (9), Others (8), Senior High school / Grade 12 (5), School level (5), High school / Grade 10 (4), Vocational courses (3).
Supported Tasks
- Question answering and multiple-choice question answering
- Multimodal (text + image) reasoning, for rows that reference images
- Instruction tuning and evaluation of LLMs on domain-specific, exam-style questions
- Explanation generation and chain-of-thought style supervision, using the
explanationfield - Retrieval-augmented generation (RAG) evaluation, using
passageas context where present
Dataset Structure
Data fields
| Field | Type | Description |
|---|---|---|
id |
string | Unique record ID, prefixed with the domain (e.g. finance_016). |
domain |
string | Top-level domain (6 values, see above). |
subject |
string | Subject area within the domain. |
category |
string | Finer-grained topic label (e.g. Quantitative Aptitude, Foreign Exchange Rates). Labels are free text and a few near-duplicates exist (e.g. English / English Language). |
grade |
string | Academic or exam level the question targets. |
answer_type |
string | single_select (choose one option) or input (type the answer). |
passage |
string | Optional shared context for the question. Empty string when not used (populated in 17 rows). |
question |
string | The question text. May contain HTML and LaTeX. |
options |
list[string] | Answer choices (4 or 5 items). Empty list for input questions. |
answer_index |
int32 | Zero-based index of the correct option in options. -1 for input questions. |
answer |
string | The correct answer as text. For single_select it equals options[answer_index]. |
explanation |
string | Worked explanation or rationale. May contain HTML. |
images |
list[struct] | Embedded images, each {bytes: binary, format: string}. All are PNG. Empty list when there are none. |
origin_source |
string | Provenance label for the record. |
Formatting notes
- HTML markup.
question,passage,explanationand a fewoptionscontain HTML tags (<p>,<br/>,<strong>,<img>and similar). Strip or render them depending on your use case. - LaTeX. Math is often written as
\( ... \). - Image placeholders. Where an image belongs in the text, the markup references it as
<img src="image_N" ...>.image_Nis the zero-based position in that row'simageslist, soimage_0isimages[0]. - Missing values. There are no nulls. Empty values are encoded as empty strings or empty lists.
Example record
{
"id": "education_research_and_advanced_science_001",
"domain": "Education, Research & Advanced Science",
"subject": "Engineering",
"category": "A.C Fundamentals, Circuits And Circuit Theory",
"grade": "Graduate",
"answer_type": "single_select",
"passage": "",
"question": "The purpose of a parallel circuit resonance is to magnify",
"options": ["Current", "Voltage", "Power", "Frequency"],
"answer_index": 1,
"answer": "Voltage",
"explanation": "In a parallel circuit resonance, the purpose is to <strong>magnify voltage</strong>. ...",
"images": [],
"origin_source": "EduGorilla"
}
Usage
from datasets import load_dataset
ds = load_dataset("InfoBay.AI/<your-repo-name>", split="train") # replace with the actual repo id
print(ds[0]["question"])
print(ds[0]["options"], ds[0]["answer"])
Reading the file directly:
import pandas as pd
df = pd.read_parquet("dataset.parquet")
# Multiple-choice rows only
mcq = df[df["answer_type"] == "single_select"]
# Rows that include images, and saving the first image of one row
with_images = df[df["images"].apply(len) > 0]
row = with_images.iloc[0]
open("example.png", "wb").write(row["images"][0]["bytes"])
Dataset Creation
Each record carries an origin_source label identifying where the question came from. Questions were organized into the six domains and subject/category/grade labels shown above, and each record includes a worked explanation and, where needed, embedded images.
Considerations and Limitations
- Small sample. 70 rows is not enough for training or for statistically meaningful benchmarking. Use it to assess schema and quality.
- Uneven distribution. Domains, subjects and grade levels are not balanced (Finance has twice as many rows as any other domain, and some categories appear only once).
- Exam-style content. Questions come from exam-preparation material, so style, difficulty and regional context (for example, Indian economy and banking topics) reflect that.
- Multimodal dependency. Some questions cannot be answered without the embedded images. Text-only models will not see them unless the images are supplied.
- Raw markup. HTML and LaTeX are preserved as-is and may need cleaning before use.
- Not professional advice. Medical, nursing, legal-adjacent and financial items are for educational and evaluation purposes only and should not be used for real-world clinical or financial decisions.
- Answer verification. Verify answers and explanations before using them as ground truth in high-stakes settings.
License
Released under CC BY 4.0. Attribution to the dataset provider is required.
Contact
Provided by InfoBay.AI, a provider of large-scale multimodal datasets and AI data services for training and evaluation. For access to the full dataset, custom subsets, licensing questions or data services, reach out through the InfoBay.AI website, LinkedIn, Hugging Face or GitHub pages.
Citation
@dataset{infobayai_multidomain_qa,
title = {Multi-Domain Question Answering Dataset},
author = {InfoBay.AI},
year = {2026},
note = {70-row sample}
}
- Downloads last month
- 67