INDEX

Explanations

instances of the number "0" as a significant feature in the text

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

code comment authors and versions

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

сылкі

-0.45

хіво

-0.41

inä

-0.35

 benefic

-0.34

 الرياضيه

-0.34

cá

-0.33

govina

-0.33

 anzuzeigen

-0.33

 Italijanski

-0.33

UrlEncoded

-0.33

POSITIVE LOGITS

ViewImports

0.63

GEBURTS

0.60

Bromoform

0.52

intios

0.50

rrggbb

0.50

 saveiro

0.47

VersionUID

0.47

IContainer

0.47

ArgsConstructor

0.46

ReusableCell

0.46

Activations Density 0.001%

instances of the number "0" as a significant feature in the text

code comment authors and versions

No Comments

No Known Activations