terminator

Name: terminator
Availability: InStock
Author: NVIDIA-NeMo

Analyze textual diagnostic reports to detect qualitative plateaus in optimization workflows.

58|16|Updated May 14, 2026

Search Tags:#optimization #analysis #terminator #decision-making #tuning #plateau

Ready to run

npx skills add https://github.com/NVIDIA-NeMo/nemo-platform --skill terminator

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: terminator
Source: https://github.com/NVIDIA-NeMo/nemo-platform/tree/main/plugins/nemo-experimentalist/src/nemo_experimentalist_plugin/skills/terminator
Command: npx skills add https://github.com/NVIDIA-NeMo/nemo-platform --skill terminator

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill addresses the challenge of determining when an optimization process has reached a point of diminishing returns that numeric metrics alone fail to capture.

Core Features & Use Cases

Qualitative Plateau Detection: Analyzes round-by-round textual analysis to identify stagnation.
Evidence-Based Decision Making: Flags recurring root causes, architectural convergence, or diagnostic degradation.
Use Case: During an automated model tuning process, use this Skill to decide whether to continue or terminate the experiment when numeric scores remain stable but the underlying mutations show no meaningful architectural progress.

Quick Start

Invoke the terminator skill to evaluate the latest round analysis and determine if the optimization process should be stopped.

Frequently Asked Questions about terminator

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I detect qualitative plateaus during model tuning when numeric metrics remain stable?▼

Detect qualitative plateaus by analyzing round-by-round textual diagnostic reports to identify stagnation in optimization progress. This approach flags recurring root causes and architectural convergence that numeric scores alone miss, helping determine if further tuning yields diminishing returns.

When do I need automated stop-checks for an iterative optimization workflow?▼

Automated stop-checks are needed when an iterative optimization process reaches diminishing returns that numeric metrics fail to capture. Use them to evaluate round-specific analysis logs against convergence criteria and decide whether to continue or terminate the experiment.

How do I set up qualitative convergence criteria to stop an automated tuning process?▼

Set up qualitative convergence criteria by providing round-specific textual analysis logs for comparative assessment. The system evaluates these diagnostic reports to detect evidence of architectural convergence, recurring root causes, or diagnostic degradation across iterations.

Can I evaluate optimization progress without relying solely on quantitative benchmark scores?▼

Yes, evaluate optimization progress by analyzing textual diagnostic reports to detect qualitative plateaus. This evidence-based method examines underlying mutations and architectural progress, providing decision-making support when numeric scores remain stable but lack meaningful structural changes.

What are the limitations of using textual analysis logs for plateau detection in decision-making?▼

The primary limitation is the dependency on access to round-specific analysis logs to perform comparative assessment. Without consistent textual diagnostic reports documenting architectural changes and root causes across iterations, qualitative plateau detection cannot be executed accurately.

What is the best way to determine if further optimization rounds will yield diminishing returns?▼

The best way to determine diminishing returns is comparing round-by-round textual analysis to detect qualitative plateaus. This automated stop-check evaluates diagnostic degradation and architectural convergence against established criteria to flag when an optimization process should terminate.