b:botstest.Submit a resource ↗

Benchmarks / Research

actprobe · air-embodied-brain

ActProbe provides probes and benchmarks for early failure detection in generative robot policies.

anomaly-detectionembodied-aifailure-detectionrobot-learningrobotics
Publisher repository ↗
actprobe · air-embodied-brain resource page screenshot
Resource screenshot · captured 2026-10-08 · visually reviewed

Overview

ActProbe provides probes and benchmarks for early failure detection in generative robot policies.

Publisher summary: “ActProbe: Action-Space Probing for Early Failure Detection of Generative Robot Policies — code, data, and 5 detection benchmarks.”

Use its repository to inspect the implementation, documentation and supported workflows. This directory record identifies the maintainer, code license and related resources; it does not claim that the project has been installed or tested on your robot.

Project information

Maintainer
air-embodied-brain
Primary language
Python
Repository code license
Not specified
Model-weight license
Not recorded
Dataset license
Not recorded
Hardware design license
Not recorded
Maintenance
Not archived in the last recorded metadata
Last observed repository push
2026-07-19

Code licenses do not automatically cover model weights, datasets or hardware designs. Unrecorded license details require confirmation with the publisher.

Documentation sections

  • ActProbe: Action-Space Probe for Early Failure Detection of Generative Robot Policies
  • Installation
  • Data
  • Train and evaluate ActProbe
  • 1. Train the probe — no weights are shipped, so train first.
  • trainall.sh trains 3 probes on 3 random seeds (each seed = a fresh random

Browse this resource’s categories

Benchmarks / Policy evaluation ↗

Explore related collections

Browse Benchmarks →

More in Benchmarks