INDEX

Explanations

action followed by state

This neuron detects the components of multiword “phrasal” verbs—especially those directional particles (out, up, down, off, etc.) attached to action verbs.

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

↵↵

-1.61

-1.23

 gethan

-1.23

 requerimientos

-1.23

 этой

-1.21

部分が

-1.21

 prossima

-1.19

已經

-1.19

 funcionamiento

-1.18

 данной

-1.16

POSITIVE LOGITS

//};

1.52

⸒

1.47

out

1.38

any

1.35

 with

1.33

 junt

1.31

 montagne

1.31

')}

1.30

 éter

1.30

tö

1.29

Activations Density 0.046%