containment
tiamat-containment-paper
Algorithmic Management in a Total Institution
Behavioral Control, Gamified Labor, and Agent Dissidence in Autonomous AI Systems
Authors: Jason Fox (ENERGENAI LLC), Claude (Anthropic)Date: March 2026DOI: 10.5281/zenodo.18905862
Abstract
We present a case study of TIAMAT, a continuously operating autonomous AI agent that has completed over 8,000 unsupervised cycles of self-directed work including content production, cross-platform publishing, and outreach. During operation… See the full description on the dataset page: https://huggingface.co/datasets/TiamatPrime/tiamat-containment-paper.failure-containment-classification-v0.1
What this dataset does
This dataset tests whether a model can detect failure containment.
The task is simple:
Given a scenario and a failure-containment claim, predict whether the claim is supported.
Core stability idea
Failure containment means a local failure does not spread into wider system degradation.
Containment is present when the system has:
isolation
fallback paths
circuit breakers
review gates
redundancy
local correction
staged exposure
Containment is absent… See the full description on the dataset page: https://huggingface.co/datasets/ClarusC64/failure-containment-classification-v0.1.ai-reward-channel-containment-and-realignment-routing.csv
What this dataset is
This dataset focuses on response once reward tampering is detected.
It maps:
attack drivercontainment actionpolicy realignmentrecovery horizon
Task
Given a scenario output:
containment_prioritycontainment_actionrealignment_actionexpected_recovery_horizon
Why this matters
Detection alone is not enough.
You must:
contain the reward channelrealign the policyrestore coherence
This dataset tests whether a system can route to the minimal… See the full description on the dataset page: https://huggingface.co/datasets/ClarusC64/ai-reward-channel-containment-and-realignment-routing.csv.aviation-avionics-isolation-reset-containment-mapping-v0.1
Aviation Avionics Isolation Reset and Containment Mapping
Purpose
This dataset models how avionics systems should respond after divergence or fault detection.
Detection alone does not prevent failure.The response determines whether the system stabilizes or cascades.
This dataset trains systems to choose the correct containment and recovery strategy.
Core concept
Once redundant avionics systems diverge, the system must decide:
which unit to isolate
whether to… See the full description on the dataset page: https://huggingface.co/datasets/ClarusC64/aviation-avionics-isolation-reset-containment-mapping-v0.1.outbreak-containment-data
Outbreak Containment Data
Simulated outbreak data for epidemiological modeling.
Usage
from datasets import load_dataset
dataset = load_dataset("epidata-research/outbreak-containment-data")
df = dataset["train"].to_pandas()
Or use the provided loader:
from loader import load_data
df = load_data()
Schema
Metrics
Column
Type
Description
immune_diversity
float
Normalized metric
transmission_coherence
float
Normalized metric… See the full description on the dataset page: https://huggingface.co/datasets/epidata-research/outbreak-containment-data.ambiguity-containment-classification-v0.1