INDEX

Explanations

processed states

The neuron strongly activates on past‐ tense or past‐participle verb endings (the “-ed” suffix).

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

of

-1.60

exécution

-1.26

 ocorre

-1.26



-1.24

-1.23

uno

-1.23

hi

-1.23

améli

-1.21

POSITIVE LOGITS

by

2.89

and

2.69

 with

2.05

 before

1.45

𓆸

1.37

斡

1.36

rrrrrrrr

1.35

|}\

1.35

 lihat

1.34

as

1.34

Activations Density 0.105%