INDEX

Explanations

mathematical and graphical representations of relationships and properties within structured data

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

graph vertices, nodes, or elements

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

 זו

-0.37

-0.35

 сем

-0.33

 זאת

-0.33

Tak

-0.33

 similar

-0.33

MQ

-0.32

Ecco

-0.32

ılığı

-0.32

any

-0.31

POSITIVE LOGITS

CloseOperation

0.65

tonode

0.60

 المعيارى

0.58

DispatchToProps

0.57

Geplaatst

0.56

UpInside

0.55

 geſch

0.54

sidemargin

0.54

RenderAtEndOf

0.54

 שוליים

0.53

Activations Density 0.203%

mathematical and graphical representations of relationships and properties within structured data

graph vertices, nodes, or elements

No Comments

No Known Activations