INDEX

Explanations

phrases indicating potential conditions or options regarding processes or decisions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 20-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_20/width_16k/average_l0_68

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

дю

-0.34

 Serbia

-0.33

 Colombia

-0.32

<?

-0.32

MFC

-0.31

 Sache

-0.31

 Ecuador

-0.31

 Doch

-0.31

 that

-0.30

 Krim

-0.30

POSITIVE LOGITS

 nakalista

0.71

AnchorStyles

0.59

argout

0.58

 akár

0.58

 للاسماء

0.57

 sogar

0.56

GIVEREF

0.56

 ब्रेकडाउन

0.55

bootstrapcdn

0.55

 zelfs

0.53

Activations Density 0.286%

phrases indicating potential conditions or options regarding processes or decisions

No Comments

No Known Activations