INDEX

Explanations

describing an issue

This neuron detects occurrences of the word “issue.”

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

].

-3.02

는

-2.61

 arbeitete

-2.59

鸶

-2.53

ineno

-2.52

敝

-2.48

।

-2.47

."

-2.44

 этого

-2.41

糶

-2.41

POSITIVE LOGITS

2.38

鈇

2.23

 också

2.23

2.22

THE

2.17

2.14

 betyder

2.13

 находя

2.11

 virker

2.11

 continúas

2.09

Activations Density 0.050%