INDEX

Explanations

terms and phrases related to actions or processes involving control and regulation

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

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

;++

-0.33

 واتس

-0.33

tag

-0.33

honneur

-0.32

whis

-0.31

runOnUiThread

-0.31

Duty

-0.31

ChildScrollView

-0.31

 Brunnen

-0.31

Now

-0.30

POSITIVE LOGITS

 kasarigan

0.86

 OFDb

0.75

featureID

0.67

 EconPapers

0.65

 autorytatywna

0.65

SharedDtor

0.65

NameInMap

0.65

httphttps

0.64

 cherchés

0.63

 snippetHide

0.61

Activations Density 0.049%

terms and phrases related to actions or processes involving control and regulation

No Comments

No Known Activations