INDEX

Explanations

`for` loop `do` statement

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

the keyword "do" in programming code, particularly in loops and block definitions.

oai_token-act-pair · claude-4-5-sonnet

Triggered by @bfgcn7frf5

The neuron fires on occurrences of the keyword “do” that begin loops or code blocks.

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

Negative Logits

じゃなくて

-1.58

-1.52

-1.48

},

-1.45

\\

-1.43

𓐍

-1.43

Creed

-1.41

was

-1.40

-1.32

POSITIVE LOGITS

</h2>

1.77

Cependant

1.56

."));

1.54

.")]

1.52

stander

1.52

It

1.51

However

1.49

疒

1.49

殼

1.45

 Aktivität

1.43

Activations Density 0.027%

`for` loop `do` statement

the keyword "do" in programming code, particularly in loops and block definitions.

The neuron fires on occurrences of the keyword “do” that begin loops or code blocks.

No Known Activations

`for` loop `do` statement

the keyword "do" in programming code, particularly in loops and block definitions.

The neuron fires on occurrences of the keyword “do” that begin loops or code blocks.

No Known Activations