INDEX

Explanations

location or recent time

This neuron detects precise spatial or directional location cues in text—especially occurrences of “just” when specifying how far or in what direction something lies.

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

 just

-1.89

まずは

-1.23

Just

-1.14

ちょうど

-1.10

 juste

-1.08

just

-0.98

 Just

-0.98

 deportiva

-0.97

JUST

-0.95

isier

-0.94

POSITIVE LOGITS

 outside

1.61

 after

1.54

 before

1.45

 above

1.31

 below

1.30

 inside

1.21

 past

1.17

 north

1.09

 west

1.09

 behind

1.05

Activations Density 0.020%