INDEX

Explanations

instances of strong emphasis or importance in the document

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

New Auto-Interp

Top Features by Cosine Similarity

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

Configuration

google/gemma-scope-2b-pt-res/layer_20/width_16k/average_l0_71

Prompts (Dashboard)

36,864 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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

omiast

-0.62

큼

-0.61

notifyListeners

-0.57

-0.53

 arşivlendi

-0.53

GMENT

-0.52

 FetchType

-0.51

)~

-0.51

 useStyles

-0.51

AddTo

-0.50

POSITIVE LOGITS

 évêque

0.65

لينكات

0.63

throw

0.61

 fallu

0.60

 figliu

0.59

Autoritní

0.58

 carattere

0.57

kano

0.57

 stratégique

0.57

 déploy

0.57

Activations Density 0.163%

instances of strong emphasis or importance in the document

No Comments

No Known Activations