INDEX

Explanations

Immunohistochemical study

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

list comprehension patterns with a "for item in collection" structure.

oai_token-act-pair · claude-3-7-sonnet-20250219 Triggered by @neilrathi

The neuron detects loop iteration constructs, i.e. the “for … in …” pattern (highlighting the loop variable and the “in” keyword).

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

 بسیار

-1.34

ystema

-1.33

 while

-1.33

GRAFIA

-1.30

ρίου

-1.26

 much

-1.26

GRAFÍA

-1.22

ḃ

-1.20

兼

-1.20

 pengurus

-1.18

POSITIVE LOGITS

was

1.78

1.63

<bos>

1.60

are

1.60

1.57

his

1.53

这种

1.50

is

1.47

1.44

Activations Density 0.004%

Immunohistochemical study

list comprehension patterns with a "for item in collection" structure.

The neuron detects loop iteration constructs, i.e. the “for … in …” pattern (highlighting the loop variable and the “in” keyword).

No Comments

No Known Activations

Immunohistochemical study

list comprehension patterns with a "for item in collection" structure.

The neuron detects loop iteration constructs, i.e. the “for … in …” pattern (highlighting the loop variable and the “in” keyword).

No Comments

No Known Activations