INDEX

Explanations

class, program, or field names

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

numeric digits, especially in sequences or identifiers within code and structured data.

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

The neuron fires on tokens that include numerals (e.g. version numbers, class or ID names, numeric codes).

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

-1.77

</h2>

-1.66

-1.54

ől

-1.34

PUESTA

-1.33

̣i

-1.32

);

-1.28

are

-1.27

–

-1.20

 even

-1.19

POSITIVE LOGITS

meanor

1.48

 большин

1.48

的事情

1.48

棡

1.47

UNITIES

1.47

1.41

 umgeben

1.40

 ähnlich

1.40

 którzy

1.32

퐿

1.31

Activations Density 0.002%

class, program, or field names

numeric digits, especially in sequences or identifiers within code and structured data.

The neuron fires on tokens that include numerals (e.g. version numbers, class or ID names, numeric codes).

No Comments

No Known Activations

class, program, or field names

numeric digits, especially in sequences or identifiers within code and structured data.

The neuron fires on tokens that include numerals (e.g. version numbers, class or ID names, numeric codes).

No Comments

No Known Activations