---
name: tai-ch139-top-average-attention-links
description: 'Apply chapter 139 of Testing AI, Top Average Attention Links, as a workflow for evaluating AI and non-deterministic systems. Use for test planning, eval design, quality review, release evidence, examples, or coaching related to top average attention links.'
---

# Top Average Attention Links

Skill name: `tai-ch139-top-average-attention-links`

Based on **Testing AI: Engineering Confidence in Non-Deterministic Systems** by **Jason Arbon**.

## Purpose

Average attention links can reveal which tokens repeatedly influence other tokens, but they are
debugging signals, not proof of reasoning.

## Use This Workflow

- Identify the AI behavior or release decision being evaluated.
- Define realistic cases, slices, unacceptable outcomes, and evidence needed for confidence.
- Choose measurements that match the risk: rubric scores, samples, intervals, traces, human review, deterministic checks, or production monitors.
- Report uncertainty, severe failures, and decision impact instead of only a pass/fail result.

## Key Guidance

Average attention links summarize which tokens tend to attend to which other tokens across heads
or layers. They reduce a huge attention tensor into a readable graph.

## Apply The Approach

Create representative cases, score them with explicit criteria, review severe failures separately, report uncertainty, and connect the evidence to a concrete decision.

## Deeper Guidance

At scale, average attention links should be compared across model versions, prompt templates,
and data slices. Watch for attention moving away from policy text, citations, constraints, or
high-risk entities.