INDEX

Explanations

words associated with research and studies

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 18-gemmascope-mlp-65k

Configuration

google/gemma-scope-2b-pt-mlp/layer_18/width_65k/average_l0_88

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

65,536

Data Type

float32

Hook Name

blocks.18.hook_mlp_out

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

 basic

-0.43

ál

-0.42

 autorytatywna

-0.42

 hard

-0.42

ymo

-0.41

.")

-0.41

&'

-0.41

Save

-0.40

manes

-0.40

hard

-0.40

POSITIVE LOGITS

webElementXpaths

0.70

paravant

0.67

 CURIAM

0.66

 مشين

0.64

TargetException

0.62

UserScript

0.62

 MonoBehaviour

0.59

LookAnd

0.57

각주

0.57

Примечания

0.57

Activations Density 0.058%

words associated with research and studies

No Comments

No Known Activations