sldbench/sldbench

SLDBench: first benchmark for scaling-law discovery — tasks curated from LLM training experiments where agents must autonomously fit and extrapolate scaling laws.

← Back to Registry

Run this task

CLI:

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

Python:

from inspect_ai import eval
from inspect_harbor import sldbench

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

Dataset information

Harbor registry	sldbench/sldbench
Inspect task	sldbench
Latest digest	sha256:369ce8a4825cae7cfb75ef0f5886f3081f072123fd37630f6d6aeef8dec46089
Samples	8
Paper	arxiv
Source	https://github.com/linhaowei1/SLD

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