INDEX

Explanations

numerical data and mathematical notations

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

in vitro recordings

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

enderror

-0.32

}{*}{}

-0.30

IContainer

-0.28

izarse

-0.28

 originaux

-0.27

klü

-0.27

้อง

-0.27

mlu

-0.27

interopRequire

-0.27

 lettres

-0.25

POSITIVE LOGITS

OGND

0.91

msgTypes

0.73

featureID

0.72

 ProtoMessage

0.67

 مشين

0.65

 censiti

0.60

पया

0.59

Abitanti

0.59

 estekak

0.59

 userSchema

0.56

Activations Density 0.037%

numerical data and mathematical notations

in vitro recordings

No Comments

No Known Activations