INDEX

Explanations

Q explanation syntax

The neuron fires on numeric literal tokens (digit sequences), i.e. it detects numbers.

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

studium

-0.97

dfs

-0.77

cektir

-0.75

referer

-0.74

 emulsi

-0.73

 clásico

-0.73

 cordoba

-0.72

uls

-0.72

','',

-0.71

kys

-0.71

POSITIVE LOGITS

amata

0.85

Vertices

0.84

SAS

0.83

 Matter

0.78

 Body

0.76

vesen

0.74

 particle

0.73

 decom

0.73

 body

0.73

 Edinburgh

0.72

Activations Density 0.004%