implementation detail · not yet filed
benchmark-instrumented target code for agent optimization
- source
- 1
- model
- 1
- lab adopt it
- 1
- strongest
- used
How sources treat it
One count per evidence span, weakest treatment to strongest.
used 1
Documented in
Evidence
1 span quoted from the sources, strongest treatment first.
For this task, we instrumented the harness with benchmarks so the model could see exactly where the time and memory went.
usedsoftware implementationin agent harnessPoolside