INDEX

Explanations

references to a particular group or subset of items

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 4-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_4/width_16k/average_l0_124

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.4.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

hek

-0.57

heide

-0.56

ôr

-0.56

무

-0.52

か

-0.51

enterOuterAlt

-0.51

lé

-0.51

Doc

-0.51

DOCUMENT

-0.51

iern

-0.50

POSITIVE LOGITS

setVerticalGroup

0.67

CHOR

0.65

0.64

ngdoc

0.64

 defaultstate

0.64

getSystemService

0.63

createStatement

0.63

IVersion

0.63

parsedMessage

0.63

 Marlin

0.62

Activations Density 0.126%

references to a particular group or subset of items

No Comments

No Known Activations