INDEX

Explanations

pending actions

The neuron flags infinitive constructions that express anticipation of a forthcoming action—especially “waiting to …” and similar “to [verb]” phrases indicating something pending.

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

还需要

-2.28

<bos>

-1.98

为什么会

-1.84

需要注意

-1.83

紧接着

-1.81

也會

-1.73

自然是

-1.72

接下来的

-1.70

 then

-1.68

 Could

-1.66

POSITIVE LOGITS

鲀

2.08

༔

2.06

they

1.62

huriyet

1.57

 sosok

1.57

 selepas

1.56

thenburg

1.55

Ᏺ

1.52

nsito

1.52

 amore

1.49

Activations Density 0.004%