INDEX

Explanations

phone numbers

The neuron strongly activates on digit tokens—especially in phone‐number–style numeric sequences.

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

 škoda

-1.37

noo

-1.26

outils

-1.23



-1.22

уме

-1.20

乄

-1.18

 Tengo

-1.17

drück

-1.17

 passend

-1.17

 Según

-1.16

POSITIVE LOGITS

–

1.45

 data

1.33

可以用

1.20

 doing

1.17

1.13

——”

1.12

 werdet

1.11

one

1.10

1.09

Activations Density 0.098%