INDEX

Explanations

think of, about, along

This neuron fires on occurrences of the word “Think,” especially when it begins a phrase.

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

หลาย

-1.67

enangan

-1.62

สำหรับ

-1.62

会長

-1.61

тации

-1.57

 president

-1.56

tention

-1.55



-1.55

ATTLE

-1.55

 และ

-1.52

POSITIVE LOGITS

er

2.03

1.82

 your

1.79

硨

1.73

You

1.72

 beschä

1.71

別に

1.63

 hablaba

1.63

后续

1.58

’,

1.57

Activations Density 0.005%