INDEX

Explanations

dimensions and measurements related to layout and design

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

pixel measurements

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.20.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

stylers

-0.58

tonode

-0.56

endphp

-0.55

лтамалар

-0.55

newArrayList

-0.50

ViewImports

-0.49

IsContent

-0.48

qtype

-0.47

StructEnd

-0.46

 déput

-0.45

POSITIVE LOGITS

 compact

0.45

 ModelExpression

0.43

compact

0.42

FormTagHelper

0.41

ThroughAttribute

0.39

 compacto

0.36

Compact

0.36

ondissement

0.32

 Gerät

0.32

 autorytatywna

0.32

Activations Density 0.006%

dimensions and measurements related to layout and design

pixel measurements

No Comments

No Known Activations