What it does
Tests whether filtering coding-agent tool results through Jev improves task success, context size, or cost against an unfiltered arm.
Benchmarks & research
Only you can see your notes.
Screenshot unavailable. Open the experiment ↗
Tests whether filtering coding-agent tool results through Jev improves task success, context size, or cost against an unfiltered arm.
Does your AI helper need all the text it reads to fix a bug? A developer can use this project to compare the same task with some text hidden and with nothing hidden. Check the finished work and recorded cost. Hidden text can be requested again, so the saving may disappear. This version has not shown a benefit worth its extra steps.