INDEX

Explanations

first of two

This neuron detects structural enumeration cues—words that introduce numbered sections or listed items (e.g. “first,” “two,” “applications,” “categories,” “aims”).

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

various

-1.22

 diverses

-1.21

 various

-1.13

 verschillende

-1.09

 such

-1.09

 diversos

-1.07

ticularly

-1.04

several

-1.02

ítimo

-1.01

digde

-1.01

POSITIVE LOGITS

 first

1.99

一是

1.62

 BOTH

1.61

Both

1.58

 Kedua

1.57

 both

1.52

 einerseits

1.36

 firstly

1.33

 Both

1.27

 первом

1.24

Activations Density 0.048%