INDEX

Explanations

reported speech and expressions conveying someone's statements or thoughts

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

 للاسماء

-0.78

featureID

-0.72

uxxxx

-0.71

parsedMessage

-0.67

 noDo

-0.66

 ब्रेकडाउन

-0.63

 surla

-0.62

 المعيارى

-0.61

 autorytatywna

-0.61

 kasarigan

-0.61

POSITIVE LOGITS

Menurutnya

0.35

saraba

0.33

esModule

0.33

 scanf

0.30

 aldri

0.29

 feelings

0.28

zzá

0.28

allowNull

0.27

 said

0.27

 biggest

0.27

Activations Density 0.012%

reported speech and expressions conveying someone's statements or thoughts

No Comments

No Known Activations