INDEX

Explanations

elements related to navigation or structural components in coding or technical documents

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

 betweenstory

-0.57

 fourrure

-0.55

 poignée

-0.45

 personnalité

-0.45

dhury

-0.43

 fumée

-0.43

 Behör

-0.41

Marcia

-0.39

 buurt

-0.38

SuspendLayout

-0.38

POSITIVE LOGITS

 nahilalakip

0.78

 Roskov

0.63

 autorytatywna

0.60

ViewFeatures

0.57

Autoritní

0.57

لينكات

0.56

Notae

0.56

Хьажоргаш

0.52

Искәрмәләр

0.52

 cherchés

0.51

Activations Density 0.004%

elements related to navigation or structural components in coding or technical documents

No Comments

No Known Activations