gabeorlanski/slopcodebench

Coding

SlopCodeBench multi-checkpoint coding benchmark tasks converted for Harbor.

← Back to Registry

Run this task

CLI:

inspect eval inspect_harbor/gabeorlanski_slopcodebench --model openai/gpt-5

Python:

from inspect_ai import eval
from inspect_harbor import gabeorlanski_slopcodebench

eval(gabeorlanski_slopcodebench(), model="openai/gpt-5")

Dataset information

Harbor registry gabeorlanski/slopcodebench
Inspect task gabeorlanski_slopcodebench
Latest digest sha256:aec29354c19d3e762640ab6d7d3c63ba8fcf4895e98a5756aecb9157b5bb4ae0
Samples 36
Paper arxiv
Source https://github.com/SprocketLab/slop-code-bench

See Task Parameters for the parameter set shared across all Harbor tasks.