INDEX

Explanations

length of

This neuron detects explicit mentions of “length” and its associated numeric measurement value.

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

and

-2.19

 メニュー

-2.16

🇰

-1.95

 設定

-1.87

 概要

-1.86

Britannique

-1.84

 個室

-1.83

 Turquía

-1.80

 Atención

-1.80

POSITIVE LOGITS

the

2.23

 atau

1.99

1.96

1.86

1.77

“

1.73

 tatuagem

1.72

 или

1.71

 男人

1.67

with

1.66

Activations Density 0.030%