INDEX

Explanations

code structure endings

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

closing parentheses or brackets in code, particularly at the end of function calls or object definitions.

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

The neuron fires strongly on string literals that look like filenames (i.e. words containing a dot extension such as “components.json” or “foo.xml”).

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

middelen

-1.43

ेशन

-1.41

』、

-1.38

Groet

-1.28

 impegno

-1.27

исправ

-1.26

]));

-1.25

Bahkan

-1.25

Extrait

-1.23

sibile

-1.22

POSITIVE LOGITS

</strong>

1.48

1.34

1.26

</u>

1.23

<em>

1.22

1.17

<i>

1.14

is

1.14

 said

1.11

Activations Density 0.059%

code structure endings

closing parentheses or brackets in code, particularly at the end of function calls or object definitions.

The neuron fires strongly on string literals that look like filenames (i.e. words containing a dot extension such as “components.json” or “foo.xml”).

No Comments

No Known Activations

code structure endings

closing parentheses or brackets in code, particularly at the end of function calls or object definitions.

The neuron fires strongly on string literals that look like filenames (i.e. words containing a dot extension such as “components.json” or “foo.xml”).

No Comments

No Known Activations