INDEX

Explanations

k and j in code loops

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

the loop variable 'j' in for loop declarations, especially at its initialization.

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

The neuron flags numeric literals (tokens representing numbers) in code.

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

鱨

-1.29

 zakoń

-1.25

for

-1.20

zusetzen

-1.17

 Comuni

-1.17

 hubungan

-1.16

㈰

-1.16

pls

-1.14

فحه

-1.14

connecting

-1.13

POSITIVE LOGITS

 ellas

1.30

izop

1.28

 played

1.24

ILERS

1.21

 otros

1.15

enia

1.12

 другие

1.10

"";

1.09

 estoy

1.09

 corta

1.09

Activations Density 0.015%

k and j in code loops

the loop variable 'j' in for loop declarations, especially at its initialization.

The neuron flags numeric literals (tokens representing numbers) in code.

No Comments

No Known Activations

k and j in code loops

the loop variable 'j' in for loop declarations, especially at its initialization.

The neuron flags numeric literals (tokens representing numbers) in code.

No Comments

No Known Activations