INDEX

Explanations

insert

The neuron activates on numeric tokens and digit sequences (e.g. section numbers, DOIs, and other multi‐digit references).

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

Iya

-0.86

ẹo

-0.85

 zaragoza

-0.85

 partant

-0.84

 raad

-0.83

begingroup

-0.82

 derni

-0.81

création

-0.79

':'

-0.79

Другие

-0.79

POSITIVE LOGITS

 insert

3.64

insert

3.36

Insert

3.14

 Insert

2.98

INSERT

2.84

 INSERT

2.53

 inserted

2.45

 inserting

2.36

插入

2.11

 inserts

2.09

Activations Density 0.052%