INDEX

Explanations

mechanisms of or properties relating to

The neuron triggers on formal scientific nouns that label study measures or descriptive properties—terms like “features,” “characteristics,” “patterns,” “performance,” or “mechanisms.”

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

 äuß

-1.22

也可

-1.19

 angustia

-1.18

 this

-1.16

 With

-1.16

</h3>

-1.16

 tessuto

-1.15

our

-1.13

èvres

-1.11

 slight

-1.10

POSITIVE LOGITS

of

2.09

and

1.61

และ

1.19

machen

1.07

1.05

 broder

1.03

Vay

1.02

ющую

1.00

 persil

0.98

läge

0.97

Activations Density 0.138%