INDEX

Explanations

sequential events

This neuron fires on the most salient content‐bearing tokens in a passage—i.e. prominent nouns, verbs, and key transition words that carry the core narrative or descriptive information.

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

 Dlatego

-1.19

PRWEB

-0.98

and

-0.97

 sogar

-0.96

まで

-0.94

 akhir

-0.89

gefunden

-0.89

合格

-0.87

 sakin

-0.86

imai

-0.83

POSITIVE LOGITS

 then

4.03

Then

2.75

 Then

2.64

 سپس

2.44

then

2.36

 затем

2.23

その後

2.11

然后

2.09

 thereafter

2.06

 kemudian

2.05

Activations Density 0.043%