INDEX

Explanations

talk about

This neuron detects conversational or speech-related verbs (especially various forms of “talk” and “hear”) indicating discussion or interview style language.

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

を使用した

-1.72

diensten

-1.66

梣

-1.59

 voegen

-1.55

 haza

-1.55

埏

-1.55

オリン

-1.54

 zoeken

-1.53

-1.51

 brengen

-1.51

POSITIVE LOGITS

 about

3.48

of

2.94

1.77

to

1.70

↵↵

1.68

In

1.64

</strong>

1.60

1.59

：“……”

1.53

Activations Density 0.050%