INDEX

Explanations

short identifiers followed by assignment

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

variable names that are abbreviations or shortened forms of longer words, particularly in programming contexts.

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

The neuron responds to numeric literal tokens (integer and floating‐point constants) in code.

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

to

-1.66

and

-1.59

 which

-1.50

 hiç

-1.32

at

-1.32

never

-1.27

从未

-1.23

in

-1.22

but

-1.20

rapeau

-1.20

POSITIVE LOGITS

سلام

1.49

فیلم

1.48

この

1.40

guantes

1.38

phòng

1.35

avait

1.31

گوشی

1.31

ighed

1.29

Activations Density 0.099%

short identifiers followed by assignment

variable names that are abbreviations or shortened forms of longer words, particularly in programming contexts.

The neuron responds to numeric literal tokens (integer and floating‐point constants) in code.

No Comments

No Known Activations

short identifiers followed by assignment

variable names that are abbreviations or shortened forms of longer words, particularly in programming contexts.

The neuron responds to numeric literal tokens (integer and floating‐point constants) in code.

No Comments

No Known Activations