🦜🔗
Auto-Evaluator
Demo
Playground
Docs
About
github
Parameters
Number of eval questions
1
5
10
15
Chunk size
500
1000
1500
2000
Chunk overlap
0
50
100
150
Model
Split method
Embedding algorithm
Retriever
Number of chunks to retrieve
3
4
5
Grading prompt style

Instructions
Upload a file (up to 50 MB) and choose the parameters for your QA chain. This evaluator will generate a test dataset of QA pairs and grade the performance of the QA chain. You can experiment with different parameters and evaluate the performance.
Upload Text for QA Eval
Attach a file (.txt, .pdf, .doc, .docx)
Upload Test Dataset (Optional)
Single CSV containing QA pairs (shape: [question, answer])