INDEX

Explanations

when or because

The neuron detects words that signal emerging problems, mishaps, or evaluative emotional states (e.g. “arise(s),” “lapse,” “discomfiture,” “nightmare,” “happy,” etc.).

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

毎回

-1.32

 eenig

-1.29

他也

-1.24

 souvent

-1.21

 eenige

-1.20

 télécharger

-1.17

 directement

-1.12

 milliers

-1.10

fficients

-1.09

our

-1.09

POSITIVE LOGITS

 when

4.72

if

2.84

by

2.59

when

2.36

 cuando

2.22

 quando

2.09

 with

2.08

 عندما

2.06

 когда

2.06

 because

2.06

Activations Density 0.148%