INDEX

Explanations

generating, invoking, converting, parsing, predicting, building

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

opening parentheses in function or method calls and definitions in programming code.

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

This neuron activates on code identifiers that name functions or methods (i.e. verbs like generate, parse, invoke, execute, convert, etc.).

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

也不敢

-1.34

Біографія

-1.17

怎

-1.14

 rospy

-1.12

LETIN

-1.10

 utom

-1.08

 другое

-1.06

ピ

-1.06

也不能

-1.06

 القانون

-1.05

POSITIVE LOGITS

呦

1.48

 some

1.27

 všetky

1.27

🦣

1.26



1.26

 például

1.25

 žena

1.24

 segíts

1.22

痙

1.20

 erity

1.18

Activations Density 0.026%

generating, invoking, converting, parsing, predicting, building

opening parentheses in function or method calls and definitions in programming code.

This neuron activates on code identifiers that name functions or methods (i.e. verbs like generate, parse, invoke, execute, convert, etc.).

No Comments

No Known Activations

generating, invoking, converting, parsing, predicting, building

opening parentheses in function or method calls and definitions in programming code.

This neuron activates on code identifiers that name functions or methods (i.e. verbs like generate, parse, invoke, execute, convert, etc.).

No Comments

No Known Activations