INDEX

Explanations

specific terms and expressions related to measurements and data representation

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

New Auto-Interp

Configuration

google/gemma-scope-9b-pt-res/layer_20/width_131k/average_l0_114

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

 Forder

-0.32

writerow

-0.31

ătă

-0.29

 useCallback

-0.27

 snapshot

-0.27

cei

-0.26

 păr

-0.26

 exit

-0.26

DeleteCommand

-0.26

เอก

-0.25

POSITIVE LOGITS

 transfieras

0.76

 lenker

0.72

 Мексичка

0.71

Jereo

0.68

Хьажоргаш

0.67

 beginnetje

0.67

IndentedString

0.66

 Савезне

0.64

bootstrapcdn

0.63

 tartalomajánló

0.63

Activations Density 0.094%

specific terms and expressions related to measurements and data representation

No Comments

No Known Activations