INDEX

Explanations

check if defined

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

preprocessor conditionals checking if symbols are defined in code (typically `#if !defined` or `!defined()` constructs).

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

The neuron activates strongly on tokens involved in conditional existence checks (e.g. hasattr, defined), i.e. code that tests whether a function, attribute, or macro is present.

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

 mouseClicked

-1.35

 budowy

-1.30

 longer

-1.23

 vacunación

-1.21

 grossen

-1.20

 elástica

-1.20

 kwiet

-1.18

JAJA

-1.17

namientos

-1.17

reszcie

-1.16

POSITIVE LOGITS

 what

1.34

 через

1.17

What

1.13

in

1.12

what

1.10

 erne

1.10

Waga

1.08

🥂

1.07

🫡

1.06

✨✨

1.04

Activations Density 0.023%

check if defined

preprocessor conditionals checking if symbols are defined in code (typically `#if !defined` or `!defined()` constructs).

The neuron activates strongly on tokens involved in conditional existence checks (e.g. hasattr, defined), i.e. code that tests whether a function, attribute, or macro is present.

No Comments

No Known Activations

check if defined

preprocessor conditionals checking if symbols are defined in code (typically `#if !defined` or `!defined()` constructs).

The neuron activates strongly on tokens involved in conditional existence checks (e.g. hasattr, defined), i.e. code that tests whether a function, attribute, or macro is present.

No Comments

No Known Activations