INDEX

Explanations

parentheses and commas

np_acts-logits-general · gemini-2.5-flash-lite

whitespace or formatting characters, particularly those that appear between sections or separate blocks of text.

oai_token-act-pair · claude-4-5-sonnet Triggered by @bfgcn7frf5

This neuron fires on numerical data—numbers, decimals, statistics, and numeric references in the text.

oai_token-act-pair · o4-mini Triggered by @jyhe0408

New Auto-Interp

Configuration

google/gemma-scope-27b-pt-res/layer_10/width_131k

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 dépens

-1.48

 vêtements

-1.48

 jejich

-1.35

hanced

-1.35

 céramique

-1.34

 desay

-1.34

 carburant

-1.34

 goutte

-1.32

 éta

-1.27

 chêne

-1.27

POSITIVE LOGITS

and

1.88

 with

1.69

 also

1.65

 نیز

1.56

it

1.53

in

1.50

of

1.43

 Similarly

1.43

1.40

 kiện

1.32

Activations Density 0.252%

parentheses and commas

whitespace or formatting characters, particularly those that appear between sections or separate blocks of text.

This neuron fires on numerical data—numbers, decimals, statistics, and numeric references in the text.

No Comments

No Known Activations

parentheses and commas

whitespace or formatting characters, particularly those that appear between sections or separate blocks of text.

This neuron fires on numerical data—numbers, decimals, statistics, and numeric references in the text.

No Comments

No Known Activations