INDEX

Explanations

the start of sentences or paragraphs indicated by a specific token format

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 17-gemmascope-res-65k

Configuration

google/gemma-scope-2b-pt-res/layer_17/width_65k/average_l0_125

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

65,536

Data Type

float32

Hook Name

blocks.17.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

<strong>

-1.02

 estekak

-0.93

fjspx

-0.87

__':

-0.81

ValueStyle

-0.80

DockStyle

-0.78

GEBURTSDATUM

-0.75

 educativas

-0.75

 HasFactory

-0.75

__":

-0.72

POSITIVE LOGITS

</b>

0.73

〢

0.60

</u>

0.58

UCC

0.57

saurus

0.57

’”

0.56

</i>

0.53

 lowa

0.52

%");

0.52

,’’

0.52

Activations Density 0.006%

the start of sentences or paragraphs indicated by a specific token format

No Comments

No Known Activations