← All projects
Accuracy across topic categories.
How long responses took to come back.
LLM Accuracy & Response Time Study · IE 4613 · Fall 2025 · Solo
Measuring LLM accuracy and response time across 40 topics
A statistics course project that built a 400-prompt dataset to evaluate how accurate large language model responses are and how fast they come back.
400
Prompts in the dataset
40
Topics covered
Image to come
Accuracy by category chart
Accuracy by category chart
The goal
Measure how accurate an LLM's responses are and how fast they come back, across topics.
Which LLM or LLMs you tested.
What I did
- Built a dataset of 400 prompts across 40 topics.
- Scored accuracy and response time with a standardized scoring key.
- Analyzed and visualized the results by category.
- Wrote a report summarizing the findings, limitations and next steps.
The statistical methods you used.
Image to come
Response time distribution
Response time distribution
Findings
The main results, such as which categories were least accurate and how response time varied.
What I learned
One or two sentences on what you took away from the study.