INDEX

Explanations

iti

The neuron consistently activates on the character sequence “iti” wherever it appears within words.

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

In

-2.78

本事

-2.59

This

-2.41

-2.38

 have

-2.38

hasilkan

-2.33

 Notably

-2.30

Also

-2.23

For

-2.22

...,

-2.20

POSITIVE LOGITS

iti

3.66

ementara

2.86

埸

2.70

2.41

鲧

2.38

𝟬

2.34

 北市

2.31

⸫

2.28

 gjorde

2.28

2.25

Activations Density 0.004%