INDEX

Explanations

reduce, change, delay, afford, budget

The neuron primarily detects numerical tokens and numeric expressions (e.g. years, dates, dollar amounts).

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

ׇ

-1.76

།

-1.55

 ویتامین

-1.52

おはよう

-1.51

 vernis

-1.43

zufügen

-1.41

 प्रोड

-1.41

 privilégi

-1.38

 comté

-1.38

 religieux

-1.36

POSITIVE LOGITS

it

1.98

for

1.92

 will

1.90

 everything

1.86

 things

1.81

 again

1.77

 what

1.67

 this

1.54

ּוֹ

1.46

が悪い

1.45

Activations Density 0.130%