INDEX

Explanations

references to CSS styles and styling elements

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

UI elements and styling

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

tagHelperRunner

-0.78

 resultCode

-0.52

tangentMode

-0.51

Tikang

-0.49

 utafitiHapana

-0.48

 vVar

-0.44

TableHead

-0.44

ReusableCell

-0.44

ýš

-0.44

 tewas

-0.43

POSITIVE LOGITS

 Styled

0.52

Styled

0.49

 makeStyles

0.47

$_['

0.43

styled

0.43

WebElementEntity

0.43

SequentialGroup

0.42

 الرياضيه

0.41

chakra

0.41

jsxFileName

0.39

Activations Density 0.002%

references to CSS styles and styling elements

UI elements and styling

No Comments

No Known Activations