INDEX

Explanations

ordinal numbers

The neuron strongly fires on explicit numeric references—digits (e.g. “5,” “4”) and ordinals (e.g. “ninth,” “sixth,” “fourth”)—wherever counts or positions are mentioned.

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

第一次

-0.93

首次

-0.90

uckles

-0.86

 majoring

-0.84

 pertama

-0.82

ATTLE

-0.79

ニラ

-0.77

toHaveBeen

-0.77

 الملك

-0.77

初回

-0.77

POSITIVE LOGITS

 fourth

4.56

 fifth

4.50

 sixth

4.06

 third

3.78

fifth

3.77

fourth

3.61

 seventh

3.56

 ninth

3.22

 eighth

3.14

sixth

3.09

Activations Density 0.089%