INDEX

Explanations

lists about numbers, lengths, or descriptions

The neuron fires on sections of text that look like scraped metadata or boilerplate—dense, list‐like passages full of numbers, titles, dates, and other structured identifiers.

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

<unused1012>

0.61

<unused1741>

0.60

<unused626>

0.59

<unused459>

0.58

<unused1938>

0.57

<unused458>

0.57

<unused568>

0.57

<unused348>

0.57

<unused1163>

0.57

<unused1728>

0.57

POSITIVE LOGITS

0.75

0.70

0.69

0.64

0.62

0.61

0.59

</a>

0.57

Activations Density 0.922%