Measuring
6
Error analysis
Teaching Gemma to Code
Why this handbook
How agents work
1
The agent loop
2
Tools
3
Context and tokens
4
Prompts and skills
Measuring
5
Evaluation
6
Error analysis
Making it better
7
First failing test
8
Running Gemma locally
9
Fine-tuning with LoRA
Looking back
10
Results
Appendices
Experiment log
Glossary
Table of contents
6.1
What I want to understand
6.2
What I tried
6.3
What I learned
6.4
Sources
Edit this page
Report an issue
Measuring
6
Error analysis
6
Error analysis
Note
Status: not started
6.1
What I want to understand
Why did each failed task fail: wrong file, bad patch, broken tests, timeout?
Which group of failures is worth fixing first?
6.2
What I tried
6.3
What I learned
6.4
Sources
5
Evaluation
7
First failing test