Visit my Blog to get in touch
Try out the project live here
Try out OpenAI's GPT assessment pack for yourself
This simple project turns OpenAI's GPT assessment pack into an interactive quiz for humans.
The pack itself can be found here: https://github.com/openai/simple-evals
There are six evaluation methods used - but for now, I have only used the MMLU pack.
- MMLU: Measuring Massive Multitask Language Understanding
- MATH: Measuring Mathematical Problem Solving With the MATH Dataset
- GPQA: A Graduate-Level Google-Proof Q&A Benchmark
- DROP: A Reading Comprehension Benchmark Requiring Discrete Reasoning Over Paragraphs
- MGSM: Multilingual Grade School Math Benchmark (MGSM), Language Models are Multilingual Chain-of-Thought Reasoners
- HumanEval: Evaluating Large Language Models Trained on Code
First download the MMLU CSV file:
cd raw
wget https://openaipublic.blob.core.windows.net/simple-evals/mmlu.csv
Then generate the SQL file:
cd db
python mmlu.py
This generates the file db/sql/mmlu.sql
Next create the database:
cd db
mysql -u root
source schema.sql
Finally, start a web-server in the root directory.
– Pete Rai