INDEX

Explanations

programming constructs and values

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

technical programming terms related to data types and structures.

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

The neuron is triggered by programming‐related tokens—things like code identifiers, data‐type names (e.g. “union,” “array,” “struct”), filenames/extensions (e.g. “.txt,” “.dic”), and other technical terms.

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

是怎么回事

-1.75

 variedade

-1.33

 COUNTRIES

-1.28

選ば

-1.24

 arşivlendi

-1.24

ⓝ

-1.23

 both

-1.23

 FILMS

-1.21

🢒

-1.20

 تحميل

-1.19

POSITIVE LOGITS

),

1.34

 więc

1.30

ｪ

1.24

 between

1.18

Also

1.13

capacete

1.13

enseits

1.11

כך

1.10

 ALSO

1.09

 поэтому

1.09

Activations Density 0.075%

programming constructs and values

technical programming terms related to data types and structures.

The neuron is triggered by programming‐related tokens—things like code identifiers, data‐type names (e.g. “union,” “array,” “struct”), filenames/extensions (e.g. “.txt,” “.dic”), and other technical terms.

No Comments

No Known Activations

programming constructs and values

technical programming terms related to data types and structures.

The neuron is triggered by programming‐related tokens—things like code identifiers, data‐type names (e.g. “union,” “array,” “struct”), filenames/extensions (e.g. “.txt,” “.dic”), and other technical terms.

No Comments

No Known Activations