INDEX

Explanations

patterns of whitespace and programming syntax elements

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

code-related terms

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

 ujednoznacz

-0.45

 autorytatywna

-0.38

Personendaten

-0.37

 désolés

-0.29

}")]

-0.28

ecma

-0.28

primaryColor

-0.27

 cerv

-0.27

*/}

-0.26

hfill

-0.25

POSITIVE LOGITS

aarrggbb

0.67

Aholisi

0.61

SharedDtor

0.57

initState

0.56

 appelez

0.56

 viewDidLoad

0.54

WarningLevel

0.54

 initState

0.53

 للاسماء

0.53

 initComponents

0.53

Activations Density 0.004%

patterns of whitespace and programming syntax elements

code-related terms

No Comments

No Known Activations